# CleverAgents Implementation Plan

## **CRITICAL**: Execute These Rules Without Exception
- **Specification in `./docs/specification.md`**: Before begining any task, **always** review the relevant parts of the specification document to understand the fine architectural details. When there is a discrepency between the current code base and the specification document, then always assume the specification document is correct.
- **Strictly adhere to guidelines in `./CONTRIBUTING.md`**: All rules and guidelines outlined in this file must be strictly followed at all times.
- **Python implementation scope only**: Every action described here pertains to building an idiomatic Python codebase that implements the CleverAgents architecture.
- **NO BACKWARDS COMPATIBILITY**: CleverAgents is a NEW standalone project. Do NOT maintain any backwards compatibility. No migration guides, no compatibility shims, no support for old configurations or data.
- **Living document protocol**: After finishing each checklist item (and its testing sub-items), immediately append every decision, discovery, open question, or deviation to this document under the matching Notes section. This plan remains the authoritative record.
- **Single documentation surface**: Do not create auxiliary notes elsewhere unless explicitly required. All architectural updates, troubleshooting outcomes, and contextual knowledge must flow back into this markdown file.
- **Sequential discipline**: Always begin with the first unchecked item in the checklist. Do not progress until that item, its documentation update, and its testing sub-items (including any spawned remediation tasks) are fully resolved.
- **USE MODERN PYTHON TOOLING**: This is a cutting-edge Python project that must use modern build tools and workflows. NO Makefiles, NO legacy approaches, NO helper scripts. Use Hatch exclusively for project management, nox for task automation, pyproject.toml for all configuration. Commands should be Python-native (e.g., `hatch env create`, `nox -s test`) not shell scripts or make targets. All tooling must be from the current Python ecosystem (2024+). When current tooling (such as "Behave" and "Robot Framework") can be used to solve a problem, use them rather than adding new tooling, keep it simple. NO wrapper scripts - use tools directly as designed.
- **Unit + integration + asv testing mandate**: For every coding task, author or update must include asv (airspeed velocity) performance, unit and integration tests, run them, and achieve passing results before marking the task complete. Testing subtasks are non-optional.
- **Behavior-driven testing stack**: Use Behave feature suites under `features/` for unit-level and scenario tests and Robot Framework suites under `robot/` for integration and end-to-end coverage. Keep both synchronized with the code under test and document all updates in this plan.
- **Do not use pytest style unit tests**: Under no circumstances should you write pytest styled unit tests, all unit tests should be Behave based (as noted in the last bullet point), which follows the Cucumber/Gherkin style of tests as seen under `features/`, this is why there is intentionally no `tests/` folder.
- **Test execution via nox**: Run every unit, integration, Behave, Robot, and benchmark suite exclusively through the designated `nox` sessions (e.g., `nox -s unit_tests`, `nox -s integration_tests`). Do not invoke `behave`, `robot`, or similar runners directly; if a `nox` session is missing required tooling, add the dependency to the session before rerunning.
- **Failure capture**: Any Behave or Robot failure instantly spawns a new unchecked sub-item under the same step titled `Fix - <short failure summary>`. Document the failure context in the Notes section and resolve it before moving forward.
- **Traceability requirement**: When a decision impacts future work, reference the relevant functions or modules in the Notes section using the `file_path:line_number` pattern for fast navigation.
- **unit tests coverage above 97% at all times**: unit test coverage must remain above 97% at all times. Unit tests can be run with `nox -e unit_tests` and is run as part of the default test suite run with `nox`.
- **must be statically typed**: All code at all times must use statically typed typing and must pass the static check run with `nox -e typecheck` which is run as part of the default test suite with `nox`. Under no circumstances at no point should you ignore type checking, this means never turn it off in the config files, and never use inline comments to force an type checking error to be suppressed.
- **Use existing tooling**: Always prefer nox sessions over raw commands, Behave for unit tests over new frameworks, Robot for integration tests, Hatch for dependency management.
- **Mock placement rule**: ALL mocks, test doubles, and mock implementations MUST exist only in `features/` directory. Production code in `src/` and utility scripts in `scripts/` must NEVER contain mock implementations, test data, or conditional testing behavior. Use dependency injection to swap implementations during tests.
- **CRITICAL - Implementation Checklist Separation**: The "Implementation Checklist" section MUST always remain separate and be the LAST section of this document. All development notes, design decisions, progress updates, technical details, and discoveries belong in their respective phase Notes sections (e.g., Phase 0 Notes, Phase 1 Notes, Phase 2 Notes) which appear BEFORE the Implementation Checklist. Never add content after the checklist section. The checklist is for tracking what needs to be done; the Notes sections are for documenting what was done and how.

### CONTINUOUS CHECKLIST AND KNOWLEDGE STEWARDSHIP (MANDATORY)

- **Immediate documentation loop**: After completing any amount of work toward a checklist item, append the newly discovered information, assumptions, implementation notes, and open questions to this document before proceeding. No discovery, decision, or workaround may remain undocumented.
- **Dynamic checklist maintenance**: Before marking an item complete—or returning from work in progress—review all remaining checklist entries. Update their descriptions, add clarifying subtasks, and insert new entries capturing follow-on tasks, bug fixes, or future enhancements uncovered during implementation. Place new items in the phase where the work logically belongs (current or future) and note cross-phase dependencies.
- **Implementation traceability**: Record substantive code decisions (design pattern choices, module ownership, testing strategy, risk mitigations) back into the corresponding Notes section and checklist subtasks immediately after each coding/testing session.
- **Checklist integrity**: Never remove checklist items unless they are explicitly retired with documented reasoning. Mark completed work by checking the item(s) and append new tasks or restructuring bullets only when required—preserve original wording for historical traceability.
- **Enforcement**: Treat omissions as blocking bugs—if the plan falls out of sync with reality, halt work, reconcile the discrepancy here, and only then continue.

## Guiding Principles

- Keep `docs/specification.md` as the single source of truth; this plan sequences work only and stays minimal.
- Local mode is the only supported execution target through Day 30; server connectivity is stub-only (interfaces raising `NotImplementedError`) after M6.
- Implement the plan lifecycle exactly: Strategize (read-only) → Execute (sandboxed) → Apply (validation-gated merge) → applied.
- Use namespaced identifiers for actions, actors, tools, skills, projects, and automation profiles; use ULIDs only for plans, decisions, resources, and validation attachments.
- Resource-first architecture: Resource Types define physical/virtual resources, DAG constraints, and sandbox strategy; registries persist Resources, Tools, Skills, Actors, Providers, and Validations.
- Tool runtime owns mutation: tool lifecycle (discover/activate/execute/deactivate), resource binding resolution, and ChangeSet tracking drive apply/rollback.
- Decision trees and invariants are first-class; correction recomputes only affected subtrees; the Invariant Reconciliation Actor runs before strategize.
- Deliver ACMS v1 in local mode by M6 (UKO + CRP + Context Assembly Pipeline); keep details in the spec, not here.
- Configuration-first YAML for actions, actors, tools, skills, resource types, validations, and automation profiles with `${ENV_VAR}` interpolation.
- Enforce strict typing and quality gates via `nox` (Behave + Robot + ASV) with coverage >=97%.

### Core Architectural Requirements

**Scalability**: ACMS v1 (UKO/CRP/context pipeline) in local mode by M6.

**Reliability**: Sandboxed execution, tool-based change tracking, validation-gated apply, and checkpoints per spec.

**Autonomy with Control**: Decision-tree recording, invariants, automation profiles, and correction flows per spec.

**Registry-First**: Resource/Tool/Skill/Actor/Validation registries with namespaced IDs and spec-aligned lifecycles.

### Continuous Testing and Documentation Policy
- Do not mark any parent checklist item complete until **all** subordinate Code, Document, Tests tasks and any generated `Fix - ...` tasks are resolved and the associated Notes section has the latest context.
- Every time new information appears, extend the corresponding Notes section immediately with explicit references to code locations and decisions.

## Completion Criteria
The implementation concludes only when every checklist item and spawned remediation task is checked, all Notes sections contain final decisions and references, and the full Behave and Robot test suites (unit, integration, end-to-end, benchmarking, packaging, documentation) pass without outstanding failures.

## Environment Variables for Testing

All environment variables needed during testing are stored in the `.env` file in the project root. This file contains API keys and tokens for various LLM providers and services. When implementing provider integrations or any features that require external services, use the environment variable names from this file or add new ones following the same pattern.

### Current Environment Variables

| Variable Name | Service | Usage |
|--------------|---------|-------|
| `OPENROUTER_API_KEY` | OpenRouter | Access to multiple LLM models through OpenRouter API |
| `OPENAI_API_KEY` | OpenAI | Direct access to OpenAI models (GPT-3.5, GPT-4, etc.) |
| `ANTHROPIC_API_KEY` | Anthropic | Access to Claude models |
| `GOOGLE_API_KEY` | Google AI | Access to Google's API for web searches |
| `GEMINI_API_KEY` | Google Gemini | Access to Google's Gemini models |
| `HF_TOKEN` | Hugging Face | Access to Hugging Face models and datasets |

## Development Log

This section will be updated with notes about each phase/task as they are implemented as well as forward looking remarks or insights.

### Phase 0: Discovery and Requirements Elaboration (Python Tooling)

**Status: [X] COMPLETE**

All core Phase 0 discovery tasks have been successfully completed. See the Phase 0 section in the Implementation Checklist for detailed completion records.

#### Phase 0 Notes

- 2025-11-01: Implemented CLI inventory extraction tools, server endpoint extraction, data contract extraction, shell asset extraction
- 2025-11-02: Implemented environment variable extraction and implicit behavior extraction tools
- 2025-11-04: Completed workflow parity matrix generator and cloud features identification
- All discovery artifacts stored in `docs/reference/`

---

### Phase 1: Target Architecture Definition

**Status: [X] COMPLETE**

All 10 ADRs have been created and package structure established. See the Phase 1 section in the Implementation Checklist for detailed completion records.

#### Phase 1 Notes

- 2025-11-04: Completed all 10 Architecture Decision Records
- 2025-11-05: Package structure created based on ADR-001
- All ADRs documented in `docs/architecture/decisions/`
- 2026-02-11: CLI syntax updated to spec-aligned forms (action create via `--config`, plan use positional args, project create positional). Legacy `--name`/`--project` examples retained in historical notes only.
- 2026-02-11: Action config YAML baseline (spec-aligned):
  ```yaml
  name: local/example-action
  description: Example action for CLI flows
  strategy_actor: openai/gpt-4
  execution_actor: openai/gpt-4
  definition_of_done: "All steps complete"
  ```

---

### Phase 2: Runtime Foundation (Completed Work)

**Status: Substantially Complete - Transitioning to new Architecture**

The following work from the previous implementation has been completed and will be preserved/adapted:

#### Completed Infrastructure
- [X] LangChain/LangGraph dependencies and integration (ADR-011)
- [X] PlanGenerationGraph, ContextAnalysisAgent, AutoDebugGraph workflows
- [X] Memory service with EntityMemory
- [X] SQLite persistence with Alembic migrations
- [X] CLI streaming integration
- [X] Provider adapters (OpenAI, Anthropic, Google, OpenRouter)
- [X] Actor configuration system (Stage 7.5)
- [X] Test coverage at 95%

#### Phase 2 Notes (Preserved from Previous Work)

**2025-11-22**: Week 12 Complete, Phase 2 Core Functionality DONE
- CLI Streaming Integration fully implemented
- AutoDebugGraph Implementation complete
- Mock provider enhancements with configurable failure modes
- Infrastructure improvements (DebugAttempt model, repositories)
- 2026-02-11: CLI syntax refreshed to spec-aligned forms in current execution sections; legacy `--name`/`--project` examples retained only in historical notes.

**2025-12-05**: LangSmith observability integration complete
**2025-12-08**: Actor-first provider wiring complete
**2025-12-17**: Stage 7 performance optimization complete
**2026-02-02**: Stage 7.5 Actor Configuration System complete
**2026-02-05**: Stage A1 & A2 Complete - Plan and Action Domain Models
- Created `src/cleveragents/domain/models/core/plan.py` with:
  - `PlanPhase` enum (ACTION, STRATEGIZE, EXECUTE, APPLY, APPLIED)
  - `ActionState` enum (AVAILABLE, DRAFT, ARCHIVED)
  - `ProcessingState` enum (QUEUED, PROCESSING, ERRORED, COMPLETE, CANCELLED)
  - `NamespacedName` model with parse() and str() methods
  - `PlanIdentity` model with ULID validation
  - `Plan` model with full lifecycle support
  - `can_transition()` function for phase transition validation
- Created `src/cleveragents/domain/models/core/action.py` with:
  - `ActionArgument` model with parse() method for CLI argument parsing
  - `Action` model with strategy/execution actor references
  - Argument validation including type checking
- Added 52 Behave test scenarios across 2 feature files:
  - `features/plan_model.feature` (30 scenarios)
  - `features/action_model.feature` (22 scenarios)
- All new tests pass, existing tests unaffected

**2026-02-05**: Stage A3 Complete - PlanLifecycleService
- Created `src/cleveragents/application/services/plan_lifecycle_service.py` with:
  - Full plan lifecycle management (Action -> Strategize -> Execute -> Apply -> Applied)
  - Action CRUD operations (create, get, list, make_available, archive)
  - Plan creation via `use_action()` which transitions Action to Strategize
  - Phase transition methods: `execute_plan()`, `apply_plan()`
  - State management: `start_*()`, `complete_*()`, `fail_*()` for each phase
  - `cancel_plan()` for non-terminal plans
  - Custom exceptions: `InvalidPhaseTransitionError`, `ActionNotAvailableError`, `PlanNotReadyError`
  - In-memory storage (to be replaced with persistence in Stage A5)
- Added python-ulid dependency for ULID generation
- Added 29 Behave test scenarios in `features/plan_lifecycle_service.feature`
- Total new test scenarios: 81 (30 + 22 + 29)

**2026-02-05**: Stage A4 In Progress - Plan CLI Commands
- Created `src/cleveragents/cli/commands/action.py` with:
  - `agents [--data-dir PATH] [--config-path PATH] action create` - Create new action with strategy/execution actors, definition of done, arguments
  - `agents [--data-dir PATH] [--config-path PATH] action list` - List actions with filtering by namespace, state
  - `agents [--data-dir PATH] [--config-path PATH] action show` - Show action details by ID or name
  - `agents [--data-dir PATH] [--config-path PATH] action available` - Make draft action available for use
  - `agents [--data-dir PATH] [--config-path PATH] action archive` - Archive an action (soft delete)
- Extended `src/cleveragents/cli/commands/plan.py` with v3 lifecycle commands:
  - `agents [--data-dir PATH] [--config-path PATH] plan use <action> <project>` - Use action to create plan in Strategize phase (legacy `--project` retained in old notes)
  - `agents [--data-dir PATH] [--config-path PATH] plan execute [plan_id]` - Transition plan from Strategize to Execute
  - `agents [--data-dir PATH] [--config-path PATH] plan apply [plan_id]` - Transition plan from Execute to Apply
  - `agents [--data-dir PATH] [--config-path PATH] plan status [plan_id]` - Show v3 plan status and details
  - `agents [--data-dir PATH] [--config-path PATH] plan list [--phase <phase>] [--state <state>] [--project <project>] [--action <action>]` - List v3 lifecycle plans with filtering
  - `agents [--data-dir PATH] [--config-path PATH] plan cancel <plan_id>` - Cancel a non-terminal plan
- Registered action commands in CLI main.py
- Added 15 Behave test scenarios in `features/action_cli.feature`
- Total test scenarios: 96 (81 + 15)


**2026-02-09**: Task Q0.6b Complete - README.md Setup Instructions [Brent]

- Updated README.md Quick Start: added `dev` extras to `pip install`, added `scripts/setup-dev.sh` step
- Updated README.md Developing section: fixed `oxt` typo -> `nox`, added quality/security nox sessions (security_scan, dead_code, complexity, pre_commit, adr_compliance), added note about pre-commit hooks, linked to `docs/development/quality-automation.md`

**2026-02-09**: Ruff Cleanup in src/cleveragents/ - StrEnum Migration [Brent]

- Fixed all 26 ruff findings in `src/cleveragents/` (all UP042: `str, Enum` -> `StrEnum`) + 12 consequent F401 unused `Enum` imports
- Migrated 26 enum classes across 15 files from `class Foo(str, Enum)` to `class Foo(StrEnum)`
- `StrEnum` (Python 3.11+) is the modern replacement; project targets Python 3.13
- Semantic difference: `str(StrEnum.MEMBER)` returns the value (e.g., `"foo"`) rather than `"ClassName.MEMBER"` — this is the correct/intended behavior for config/JSON string enums
- Verified: no code uses `str()` on enum members in the old format; all tests pass (304 scenarios, 0 failures)
- Files: 15 domain model files + `memory_service.py` + `providers/registry.py`

**2026-02-09**: Task Q0.9 Complete - Ruff Lint Findings in features/ [Brent]

- Fixed all 200 ruff lint findings in `features/` directory -> **0 findings**
- **Config-level suppressions** (168 findings):
  - Added `per-file-ignores` in `pyproject.toml` for Behave-specific patterns:
    - `features/steps/*.py`: F811 (65 redefined `step_impl` — Behave idiom), E501 (long step decorator strings)
    - `features/mocks/*.py`, `features/environment.py`: E501
- **Manual fixes** (31 findings across 18 files):
  - 11x SIM115: `NamedTemporaryFile` refactored to use `with` context manager (`actor_cli_steps.py`, `actor_cli_run_steps.py`)
  - 4x UP028: `for/yield` -> `yield from` (google, openai, openrouter, langchain provider steps)
  - 3x SIM117: Nested `with` -> single `with` with parenthesized contexts (`plan_full_coverage_steps.py`, `plan_service_steps.py`)
  - 3x RUF005: `list + [item]` -> `[*list, item]` unpacking
  - 2x B904: Added `from exc` to `raise` inside `except` (enums, retry patterns)
  - 2x RUF012: Added `ClassVar` annotations (`vector_store_service_steps.py`)
  - 2x SIM105: `try/except/pass` -> `contextlib.suppress(Exception)`
  - 1x each: B007 (unused loop var), B018 (noqa suppression), F821 (missing `Any` import), SIM102 (collapsible if), I001 (auto-fixed unsorted import)
- **Verification**: All affected behave tests pass (155 scenarios, 0 failures)
- **Files modified**: `pyproject.toml` (config), `environment.py`, and 17 step files in `features/steps/`

**2026-02-09**: Task Q0.8 Complete - Bandit Security Findings Remediation [Brent]

- Fixed all 16 pre-existing bandit findings (2 HIGH, 3 MEDIUM, 11 LOW) -> **0 findings**
- **Security hardening** (HIGH+MEDIUM):
  - Replaced `jinja2.Environment` with `jinja2.sandbox.SandboxedEnvironment` in `yaml_template_engine.py` and `stream_router.py` — prevents template injection
  - Added `_validate_code_ast()` helper: AST-based pre-validation for `exec()` in `SimpleToolAgent` — rejects imports, `exec()`/`eval()`/`compile()`/`__import__()`/`getattr()`/`setattr()` calls, global/nonlocal statements
  - Added `_validate_lambda_ast()` helper: restricts transform `eval()` to lambda-only expressions via AST parsing
  - Suppressed `0.0.0.0` bind default (`# nosec B104`) — intentional, configurable via `CLEVERAGENTS_SERVER_HOST`
- **Code quality** (LOW):
  - Replaced 6 `assert` statements with proper `if`/`raise` (TypeError, RuntimeError, typer.BadParameter) — asserts stripped in optimized bytecode
  - Replaced `try/except/pass` with `contextlib.suppress(Exception)` (2 locations in dispose())
  - Added logging to previously-silent exception handlers (migration_runner, nodes retry loop)
  - Suppressed false positive `"token_count": 0` flagged as hardcoded password (`# nosec B105`)
- **Files modified**: `stream_router.py`, `yaml_template_engine.py`, `settings.py`, `context_service.py`, `memory_service.py`, `retry_patterns.py`, `context.py` (CLI), `plan_service.py`, `migration_runner.py`, `nodes.py`
- **Verification**: `bandit -r src/ -c pyproject.toml` → 0 findings; targeted behave tests pass; smoke tests for AST validation pass

**2026-02-09**: Stages Q0, Q1, Q2 Complete - Full Quality Automation Setup [Brent]

**Stage Q0 - Pre-commit Hooks:**

- Created `.pre-commit-config.yaml` with 12 hooks across 5 categories:
  - Branch protection: `no-commit-to-branch` (prevents commits to main)
  - General checks: `check-yaml`, `check-toml`, `check-json`, `check-merge-conflict`, `check-added-large-files`, `end-of-file-fixer`, `trailing-whitespace`, `debug-statements`
  - Ruff: `ruff-format` (auto-fix), `ruff` (lint with safe auto-fix)
  - Pyright: local system hook running type checking on `src/` only
  - Bandit: security scanning with `pyproject.toml` configuration on `src/` only
  - Vulture: dead code detection with whitelist at `vulture_whitelist.py`
  - Semgrep: custom rules in `.semgrep.yml` for eval/exec/os.system/pickle detection (graceful skip when not installed)
  - Commitizen: conventional commit message validation at commit-msg stage
- Added dev dependencies to `pyproject.toml`: `pre-commit>=3.6.0`, `bandit[toml]>=1.7.5`, `vulture>=2.10`, `radon>=6.0.1`
- Added `[tool.bandit]` and `[tool.vulture]` sections to `pyproject.toml`
- Created `vulture_whitelist.py` for false positive suppression (exc_tb, build_data)
- Created `.semgrep.yml` with 5 custom security rules
- Created `scripts/setup-dev.sh` for developer environment setup
- Added 4 new nox sessions: `pre_commit`, `security_scan`, `dead_code`, `complexity`
- **Discovery**: CI platform is Forgejo (`.forgejo/`), NOT GitHub. Stage Q1 must use Forgejo Actions, not GitHub Actions.
- **Discovery**: Pre-existing security findings in production code need remediation (see Q0.8 spawned task)
- **Discovery**: 37 source files have formatting issues, ~27 files have trailing whitespace - pre-existing debt
- **Discovery**: `features/steps/actor_cli_steps.py` has many F811 (redefined step_impl) violations - Behave pattern
- **Discovery**: Average code complexity is A (3.56) across 979 blocks - good baseline
- **Discovery**: High complexity methods identified: `LegacyDataMigrator.migrate_project_data` E(37), `ProviderRegistry._create_provider_llm` C(20), `ProviderRegistry.create_ai_provider` C(18)
- Key files: `.pre-commit-config.yaml`, `.semgrep.yml`, `vulture_whitelist.py`, `scripts/setup-dev.sh`

**Stage Q1 - CI/CD Pipeline:**

- Extended `.forgejo/workflows/ci.yml` with 3 new jobs:
  - `security`: bandit scan (JSON report + high-severity gate) + vulture dead code detection
  - `quality`: radon complexity check (grade F fails build) + JSON report
  - `coverage`: behave tests with coverage measurement, fail-under=85%, XML artifact
- Updated `docker` and `helm` jobs to depend on `security` (fail-fast on security issues)
- Created `scripts/check-quality-gates.py` aggregating: coverage, typecheck, security, dead code, complexity
- All reports uploaded as artifacts for downstream consumption

**Stage Q2 - Advanced Automation:**

- Created `.forgejo/workflows/nightly-quality.yml` for nightly quality monitoring:
  - Runs at midnight UTC (cron: "0 0 \* \* \*") + manual trigger support
  - Full lint, typecheck, security scan (all severities), dead code, complexity analysis
  - Behave tests with coverage measurement
  - Quality trend JSON generation with timestamp + metrics
  - 90-day artifact retention for trend analysis
- Created `scripts/check-adr-compliance.py` with AST-based checks for:
  - ADR-002: No threading imports in application layer
  - ADR-003: Services use constructor dependency injection
  - ADR-007: No direct SQLAlchemy usage in service layer
- Added `nox -s adr_compliance` session
- Created `.forgejo/pull_request_template.md` with quality checklist
- Created `docs/development/quality-automation.md` with full documentation:
  - Quick start, pre-commit hooks reference, CI jobs table, security scanning guide
  - Complexity monitoring grades, quality gates, troubleshooting

**New nox sessions added:** `pre_commit`, `security_scan`, `dead_code`, `complexity`, `adr_compliance`
**Total files created:** 9 new files
**Total files modified:** 3 files (pyproject.toml, noxfile.py, .forgejo/workflows/ci.yml)

**2026-02-10**: Task 10B.4 - Quality Metrics Baseline Established [Brent]

- Ran full quality suite via nox to establish current baseline:
  - **Unit Tests**: 105 features, 1613 scenarios, 7555 steps - ALL PASS
  - **Lint (ruff)**: 0 findings
  - **Typecheck (pyright)**: 0 errors, 0 warnings
  - **Security (bandit)**: 0 findings (0 HIGH, 0 MEDIUM, 0 LOW)
  - **Dead Code (vulture)**: 0 findings
  - **Complexity (radon)**: Average A (3.56), 981 blocks analyzed, no grade-F methods
    - High complexity methods to monitor: `LegacyDataMigrator.migrate_project_data` E(37), `Action.validate_arguments` C(20), `ProviderRegistry._create_provider_llm` C(20), `ProviderRegistry.create_ai_provider` C(18), `Settings.resolve_provider_defaults` C(18)
  - **Coverage**: 96% (9860 statements, 269 missing, 2852 branches, 213 branch-miss)
- Fixed pre-existing test failure: `plan_lifecycle_cli_coverage.feature` scenario "Plan lifecycle list shows project summaries" - Rich table column wrapping at narrow terminal widths caused `+1 more` text to be split across rows. Fixed by patching console width to 200 in test setup.
- Fixed missing dependency: added `langchain-anthropic>=0.2.0` to `pyproject.toml` (was imported in `src/cleveragents/providers/llm/anthropic_provider.py` but not declared)

**2026-02-10**: Task Q1.5 Complete - Branch Protection Rules Documentation [Brent]

- Created `docs/development/ci-cd.md` (224 lines) documenting:
  - Branch protection rules for `master` (required status checks, review requirements, push/force-push/deletion blocks)
  - Step-by-step Forgejo branch protection setup instructions
  - Review priority matrix (P0: architecture/security, P1: algorithms, P2: features, P3: tests/docs)
  - CI job dependency graph and quality gates summary table
  - Nightly quality monitoring reference
  - Local development workflow quick-reference
- Cross-references existing `docs/development/quality-automation.md` and `.forgejo/pull_request_template.md`
- Required CI checks documented: `lint`, `typecheck`, `security`, `quality`, `behave`, `coverage`, `build`
- Review requirement: 1 approving review, selective depth by priority matrix

**2026-02-10**: Task 10C.1 Complete - Edge Case Test Scenarios [Brent]

- Created `features/edge_case_plan_scenarios.feature` (26 scenarios, 141 steps) covering:
  - **Concurrent plan execution** (6 scenarios): Duplicate strategize/execute/apply start attempts, concurrent complete+fail on same phase, two plans from same action, concurrent transitions on different plans
  - **Resource conflict scenarios** (7 scenarios): Read-only file MODIFY failure, overlapping CREATE+MODIFY on same path, CREATE with None content, MOVE with missing source, DELETE of already-deleted file, file paths with spaces, deeply nested directory creation
  - **Validation failure chains** (6 scenarios): Multiple simultaneous argument validation failures (missing + wrong type), unknown + missing arguments combined, empty plan description rejection, ACTION phase with processing state, invalid namespace characters, invalid name characters
  - **Rollback edge cases** (7 scenarios): Partial apply failure (first change persists on disk, second unchanged), errored plan cannot restart, fail preserves error message, errored plan not terminal, cancel preserves phase, errored strategize rejects execute, failed strategize rejects complete
- Created `features/steps/edge_case_plan_steps.py` with step definitions for all 26 scenarios
- Verified no step name collisions with existing 104 feature files
- All quality gates pass: lint 0 findings, typecheck 0 errors, full suite 106 features / 1639 scenarios / 7696 steps ALL PASS

**2026-02-10**: Task 10C.4 Complete - Validation Test Fixtures [Brent]

- Created `features/validation_test_fixtures.feature` (34 scenarios, 81 steps) covering 6 validation domains:
  - **AST security validation** (`_validate_code_ast`): 12 scenarios testing import/global/nonlocal/exec/eval/compile/**import**/getattr/setattr rejection, syntax errors, and safe code acceptance
  - **Lambda AST validation** (`_validate_lambda_ast`): 4 scenarios testing valid lambda, non-lambda rejection, syntax errors, function call rejection
  - **Python content sanitization** (`_sanitize_python_content`): 4 scenarios testing passthrough, code fence stripping, docstring wrapping, irrecoverable syntax (null byte)
  - **Project model validation**: 6 scenarios testing invalid chars, slashes, exclamation, valid names, relative path resolution, empty name
  - **Change list coercion** (`_coerce_change_list`): 4 scenarios testing empty list, mixed entries, non-list, non-change entry
  - **ActionArgument parsing**: 4 scenarios testing too few parts, invalid type, invalid requirement, reserved keyword name
- Created `features/steps/validation_test_fixture_steps.py` with complete step definitions for all 34 scenarios
- Fixed step name collisions: renamed `I create a project with name` -> `I create a project fixture with name` and `the project path should be absolute` -> `the project fixture path should be absolute` to avoid conflicts with `database_integration_steps.py` and `domain_models_steps.py`
- Moved inline import (`ArgumentRequirement`, `ArgumentType`) to file-level per CONTRIBUTING.md rules
- Fixed irrecoverable syntax scenario: `"def broken("` is actually recoverable via docstring wrapping; replaced with null byte input which is truly irrecoverable
- All quality gates pass: lint 0 findings, typecheck 0 errors, full suite 107 features / 1673 scenarios / 7777 steps ALL PASS

**2026-02-12**: Task Q0-min-ci In Progress - Nox-Based PR Validation Workflow [Brent]

- Rewrote `.forgejo/workflows/ci.yml` to route ALL jobs through nox sessions instead of direct tool invocations:
  - All jobs run `pip install uv nox` then `nox -s <session>`
  - Jobs: `lint`, `typecheck`, `security`, `quality`, `unit_tests`, `integration_tests`, `coverage` (fail-under 97%), `build`, `docker`
- Updated `noxfile.py`:
  - `format` session: passes `session.posargs` through (supports `--check` from CI)
  - `coverage_report` session: rewritten from broken parallel behave mode to serial `coverage run --source=src -m behave` mode. Changed `--fail-under` from 85 to **97**. Parallel mode was broken (produced 22% coverage due to subprocess data collection issues); serial mode correctly reports coverage.
- Updated `docs/development/ci-cd.md`: coverage threshold 85→97%, updated job names, added "Nox-Based CI" section, "Local Reproduction of CI Failures" section, "Caching Notes" section
- Created `features/ci_workflow_validation.feature` + `features/steps/ci_workflow_validation_steps.py` — 11 scenarios validating CI workflow YAML structure (job→nox session mapping, Python version, dependencies)
- Created `robot/ci_nox_validation.robot` — Robot smoke test verifying `nox --list` output and CI file existence
- Created `benchmarks/ci_yaml_parse_bench.py` — ASV benchmark for CI YAML parsing
- **Coverage boost 92%→97%**: Wrote 108 new Behave scenarios covering 6 largest coverage gaps:
  - `features/yaml_engine_direct_coverage.feature` (21 scenarios) — `actor/yaml_template_engine.py` 13%→91%
  - `features/actor_config_new_coverage.feature` (30 scenarios) — `actor/config.py` 28%→93%
  - `features/actor_registry_new_coverage.feature` (14 scenarios) — `actor/registry.py` 20%→93%
  - `features/message_router_new_coverage.feature` (4 scenarios) — `langgraph/message_router.py` 29%→100%
  - `features/context_analysis_new_coverage.feature` (19 scenarios) — `agents/graphs/context_analysis.py` 73%→93%
  - `features/context_service_new_coverage.feature` (20 scenarios) — `application/services/context_service.py` 75%→92%
- **Key discovery**: Behave loads ALL step definition files globally; any `@given/@when/@then` pattern matching an existing one causes `AmbiguousStep` errors. Biggest offender: `@then('the result should be {expected}')` in `features/steps/cli_steps.py:135`. All new step names were made unique.
- **Verification**: 1673 scenarios passed (0 failed), 97% coverage (fail-under=97 passes), lint pass, typecheck pass
- **Remaining**: Branch creation, commit, push, PR, merge, branch cleanup

**2026-02-12**: Bugfix - Integration Test Failure in Robot.Actor Configuration [Brent]

- **Root cause**: `robot/ci_nox_validation.robot` ran `nox --list` during integration tests. When `pabot` executed all Robot suites in parallel, the `nox --list` stdout leaked into the sibling `robot/actor_configuration.robot` process. The `helper_actor_config.py` JSON output had the full `nox --list` text appended, causing `JSONDecodeError: Extra data` on the `V2 Actor Config Produces Provider And Graph Descriptor` test case.
- **Fix applied (3 files)**:
  - `robot/ci_nox_validation.robot`: Added `slow` tag to "Nox Lists All Required Sessions" and "CI Workflow File Exists" test cases so they are excluded from the normal `--exclude slow` integration run (the "Nox Lint Session Runs Successfully" test already had `slow`)
  - `robot/actor_configuration.robot:31`: Replaced hardcoded `.nox/integration_tests-3-13/bin/python` with `python` (follows project convention; never reach into nox's internal venvs)
  - `robot/plan_generation_graph.robot:12`: Same fix — replaced hardcoded `.nox/integration_tests-3-13/bin/python` `${PYTHON}` variable with `python`
- **Verification**: 175 integration tests passed (0 failed), 1673 unit test scenarios passed (0 failed)

**2026-02-12**: Bugfix - CI `security` Job Failure (session name mismatch) [Brent]

- **Root cause**: `.forgejo/workflows/ci.yml` referenced `nox -s security_scan` and `nox -s dead_code`, but the noxfile only had a session named `security` (which bundled bandit + vulture). No `dead_code` session existed.
- **Fix applied (2 files)**:
  - `noxfile.py`: Renamed `security` session to `security_scan` to match CI reference. Added standalone `dead_code` session running vulture directly (intentionally duplicates the vulture step inside `security_scan` for independent CI visibility). Updated `nox.options.sessions` default list accordingly.
  - `.forgejo/workflows/ci.yml`: Updated `nox -s security` to `nox -s security_scan`, restored `nox -s dead_code` step.
- All 10 CI session references now verified against `nox --list`: lint, format, typecheck, security_scan, dead_code, complexity, unit_tests, integration_tests, coverage_report, build.

**2026-02-12**: Bugfix - CI `unit_tests` Failure: Rich Console Line-Wrapping [Brent]

- **Root cause**: Rich's `Console` wraps output at 80 columns. On CI, the temp directory path is ~90 characters (`/workspace/cleveragents/cleveragents-core/.nox/unit_tests-3-13/tmp/.../existing_0.txt`), so filenames get split across lines (e.g., `e\nxisting_0.txt`). The assertion `expected_name in clean_output` fails because `existing_0.txt` doesn't appear as a contiguous substring.
- **Failing scenario**: `features/context_unit_tests.feature:43` — "Context add reports already tracked files without overflow summary"
- **Fix applied**: `features/steps/context_unit_tests_steps.py` — Both `step_check_already_tracked` (line 588) and `step_check_already_tracked_without_summary` (line 606) now collapse Rich line-wraps via `clean_output.replace("\n", "")` before checking for filenames and overflow summaries, making assertions resilient to any console width.
- **Verification**: All 67 scenarios in `context_unit_tests.feature` pass locally.

**2026-02-12**: Bugfix - CI `integration_tests` Failures (3 categories) [Brent]

- **Issue 1 — `ModuleNotFoundError: No module named 'features'`**: The `Run Python Script` keyword in `robot/database_integration.robot` hardcoded `sys.path.insert(0, '/app')` in its header. On CI, the workspace is `/workspace/cleveragents/cleveragents-core`, so `features.mocks.mock_ai_provider` could not be found.
  - **Fix**: Changed the header path to `sys.path.insert(0, '${CURDIR}/..')` which resolves relative to the robot file location. Removed redundant hardcoded `/app` from the "Service Layer Uses Repositories" test script body.
- **Issue 2 — `--load-context` help text assertion failures**: `robot/load_context_test.robot` had three test cases checking `Should Contain    ${result.stdout}    --load-context`. On CI, Rich's help output may go to stderr when no TTY is detected, causing stdout to be empty. Also, the test `Test Load Context Interactive Help Text` was duplicated (appeared twice), which is invalid in Robot Framework.
  - **Fix**: Added `stderr=STDOUT` to `Run Process` calls so all output is merged into stdout. Removed the duplicate test case.
- **Issue 3 — `FileNotFoundError: [Errno 2] No such file or directory: 'robot'`**: pabot spawns `robot` subprocesses for each test suite. On CI, the first ~16 suites succeed but subsequent ones fail with `robot` not found. This is likely a CI container process/resource exhaustion issue, not a code defect. The `--exclude discovery` flag was also missing from the `integration_tests` nox session (dropped during merge conflict resolution).
  - **Fix**: Restored `--exclude discovery` to the `integration_tests` nox session in `noxfile.py`. The `robot` not found error is a CI infrastructure issue that may resolve with fewer parallel suites; it is not reproducible locally.
- **Verification**: `robot/database_integration.robot` and `robot/load_context_test.robot` both pass locally.

**2026-02-13**: Bugfix - CI `security` Job Failure (missing `build/` directory)

- **Root cause**: The `security_scan` nox session writes bandit's JSON report to `build/bandit-report.json`, but the `build/` directory does not exist in a fresh CI checkout. Bandit cannot create intermediate directories and fails with exit code 2: `can't open 'build/bandit-report.json': [Errno 2] No such file or directory`.
- **Fix applied (1 file)**:
  - `noxfile.py:484`: Added `os.makedirs("build", exist_ok=True)` before the first bandit invocation in the `security_scan` session. The `os` module was already imported. `exist_ok=True` makes this idempotent for both CI (fresh checkout) and local dev (directory may already exist).
- **Verification**: `nox -s security_scan` passes locally (0 bandit findings, 0 vulture findings).

**2026-02-13**: Bugfix - CI `integration_tests` Failures (3 categories)

- **Issue 1 — `FileNotFoundError: [Errno 2] No such file or directory: 'robot'`** (affected ~10 suites: Plan Generation Graph, Retry Patterns, Rxpy Route Validation, and suites #25-31): After ~17 suites complete, pabot child subprocesses fail to find the `robot` binary. This is a CI container resource exhaustion issue — pabot spawns `robot` subprocesses via `subprocess.Popen`, and under high parallelism the system cannot resolve the executable.
  - **Fix** (`noxfile.py`):
    - `_pabot_parallel_args()` (line 46-49): Capped default pabot parallelism to `min(_default_processes(), 2)`. Local developers can still override via `nox -s integration_tests -- --processes 4`.
    - `integration_tests()` session (lines 372-381): Explicitly prepend the nox venv `bin/` directory to `session.env["PATH"]` so all child subprocesses inherit a correct PATH. Added a debug line printing the resolved `robot` binary path before pabot launches.
- **Issue 2 — `Robot.Load Context Test` help text assertion failures** (2 tests: `Test Load Context Help Text`, `Test Load Context Interactive Help Text`): Both tests assert `Should Contain ${result.stdout} --load-context`. The text is present in the Rich-formatted output, but ANSI escape codes embedded by Rich split the `--load-context` string, breaking the naive substring match on CI (no TTY).
  - **Fix** (`robot/load_context_test.robot:177,183`): Added `env:NO_COLOR=1` to the `Run Process` calls for both tests. This disables Rich's ANSI escape codes, producing plain text that `Should Contain` can match reliably.
- **Issue 3 — `Robot.Initial Next Command Test` exit code 3** (1 test: `Test Next Command With Null Writing Stage`): Exit code 3 is the generic `except Exception` handler in `src/cleveragents/cli/commands/actor.py:164-166`. The `!next discovery` command triggers the `ask_topic` LLM agent (OpenAI gpt-3.5-turbo in `examples/scientific_paper_writer.yaml:671-682`). Without a valid `OPENAI_API_KEY` on CI, a `ValueError` is raised, which is not a `CleverAgentsError` subclass and falls through to exit code 3.
  - **Fix** (`robot/initial_next_command_test.robot:24`): Added `slow` tag to the test case. The `integration_tests` nox session already passes `--exclude slow` to pabot, so this test is skipped on CI. It can still be run explicitly via `nox -s slow_integration_tests`.
- **Verification**: `nox -s integration_tests` passes locally — 206 tests, 206 passed, 0 failed, 0 skipped.

**2026-02-13**: Bugfix - CI `integration_tests` Failures Round 2 (pabot resource exhaustion + ANSI codes)

- **Observation**: The first round of fixes (parallelism cap to 2, PATH propagation, `NO_COLOR=1`) did not resolve the CI failures. Debug output confirmed `robot` is findable at session start (`robot at: .nox/.../bin/robot`), yet pabot still fails with `FileNotFoundError` after ~24 suites even with `--processes 2`. The `NO_COLOR=1` env var did not prevent Rich from embedding ANSI escape codes in help text output on CI.
- **Issue 1 — `robot` not found (resource exhaustion, not PATH)**:
  - **Root cause confirmed**: The debug line proves `robot` is on PATH. The `FileNotFoundError` from `subprocess.Popen` on Linux occurs when the system cannot `execve()` due to exhausted PIDs or file descriptors, not because the binary is missing. pabot spawns a new `robot` subprocess per test suite plus a PabotLib RPC server, accumulating resources until the container limit is hit.
  - **Fix** (`noxfile.py`): Replaced `pabot` with `robot` (sequential execution) in the `integration_tests` session. Removed `_pabot_parallel_args()` function (no longer needed). `robot` runs all suites in a single process — no subprocess spawning, no resource accumulation. CI values reliability over speed. The `slow_integration_tests` session remains unchanged for opt-in parallel runs via `pabot`.
  - Added resource debug output: `os.listdir('/proc/self/fd')` count and `ulimit -n`/`ulimit -u` to help diagnose future container issues.
  - Added `os.makedirs("build/reports/robot", exist_ok=True)` for the output directory.
- **Issue 2 — Load Context Test ANSI codes (NO_COLOR insufficient)**:
  - **Root cause**: `NO_COLOR=1` only disables ANSI *color* codes. Rich may still embed non-color ANSI sequences (bold `\x1b[1m`, reset `\x1b[0m`) that split the `--load-context` string, causing `Should Contain` to fail.
  - **Fix** (`robot/load_context_test.robot:177,183`): Added `env:TERM=dumb` alongside `env:NO_COLOR=1` to both help text test cases. `TERM=dumb` causes Rich to disable all terminal styling. Added `repr()` debug logging that prints the raw bytes around the `load` substring to the CI console — this will definitively reveal any remaining invisible characters if the tests still fail.
- **Verification**: `nox -s integration_tests` passes locally — 206 tests, 206 passed, 0 failed. Debug repr output shows clean text: `'│\n│    --load-context                    FILE    '`.

**2026-02-13**: Bugfix - CI `integration_tests` Failures Round 3 (venv PATH, Resource paths, hanging tests)

- **Observation**: Round 2 fixed pabot resource exhaustion by switching to sequential `robot`. This exposed 18 new test failures (previously hidden — those suites never completed under pabot). Three root causes identified.
- **Issue 1 — `python -m cleveragents` uses system Python (not venv Python)**:
  - **Root cause**: When rewriting the `integration_tests` session to use `robot` instead of `pabot`, the `session.env["PATH"]` line that prepends the venv `bin/` directory was accidentally removed. Robot Framework's `Run Process` keyword inherits the nox session environment, so without the PATH override, `python -m cleveragents` resolved to the system Python (which lacks the installed package).
  - **Fix** (`noxfile.py`): Restored `session.env["PATH"]` to prepend the venv bin directory. Consolidated debug output into a single `session.run("python", "-c", ...)` call showing robot path, CWD, and robot/ directory contents.
- **Issue 2 — Robot Framework `Resource` file resolution fails on CI**:
  - **Root cause**: Robot files used bare relative `Resource` references (e.g., `Resource v2_paths.resource`, `Resource common.resource`). Robot Framework resolves bare paths relative to the *current working directory*, not the test file's directory. When CWD differs between local and CI environments, these references break with `Resource file 'v2_paths.resource' does not exist`.
  - **Fix** (30 robot files): Converted all bare `Resource` references to use `${CURDIR}/` absolute paths (e.g., `Resource ${CURDIR}/v2_paths.resource`). Exception: `robot/core_cli_commands.robot` retains bare `Resource discovery_common.resource` because that file intentionally does not exist (tests using it are excluded via `--exclude discovery`).
- **Issue 3 — `rxpy_route_validation.robot` tests hang indefinitely**:
  - **Root cause**: Several `Run Process` calls invoking `python -m cleveragents actor run` lacked `timeout` parameters. Tests that expect the command to succeed (LangGraph routes, allowed RxPY routes) start actual actor runs that never terminate, causing the entire test suite to hang.
  - **Fix** (`robot/rxpy_route_validation.robot`): Added `timeout=30s` to all 9 `Run Process` calls. Tagged `Test Run Command With LangGraph Routes Succeeds` and `Test Run Command Accepts RxPY Routes When Allowed` as `slow` (these require runtime infrastructure unavailable on CI). Also fixed a bug where `Test Run Command With LangGraph Routes Succeeds` was using `${RXPY_CONFIG}` instead of `${LANGGRAPH_CONFIG}`, and removed conflicting duplicate `stderr`/`stdout` parameters.
- **Verification**: `nox -s integration_tests` passes locally — 204 tests, 204 passed, 0 failed (2 additional tests now excluded via `slow` tag: the 2 RxPY tests that require running actors).

**2026-02-13**: Bugfix - CI `integration_tests` Failures Round 4 (duplicate Settings blocks, venv Python injection, CI debug)

- **Observation**: Round 3 fixes (commit `e8aa5ac`) passed locally (204/204) but CI still showed 18 failures across 6 files. Two distinct failure signatures: (A) `Resource file '/workspace/.../robot/v2_paths.resource' does not exist` for files that alphabetically follow `sandbox_integration.robot`, despite the directory listing confirming both `.resource` files are present; (B) `/usr/local/bin/python: No module named cleveragents` — system Python used instead of venv Python.
- **Issue 1 — Duplicate `*** Settings ***` blocks cause resource import failures on CI**:
  - **Root cause**: The Round 3 sed-based `Resource` path conversion created a *second* `*** Settings ***` block in 14 robot files (the `Resource ${CURDIR}/...` line was appended in a new block rather than inserted into the existing one). While Robot Framework 7.4.1 nominally merges duplicate sections, this non-standard structure caused intermittent resource import failures on the CI container (Forgejo runner with `python:3.13-slim`). Locally, the same files parsed correctly — suggesting an RF parsing race or filesystem interaction specific to the CI overlay filesystem.
  - **Fix** (14 robot files): Merged all duplicate `*** Settings ***` blocks into a single block per file, combining Documentation, Library, Resource, Suite Setup, and Suite Teardown into the canonical first block. Removed duplicate `Resource` imports (2 files had triple Settings blocks with duplicated imports). Added blank line separators between sections.
- **Issue 2 — Bare `python` in `Run Process` calls resolves to system Python on CI**:
  - **Root cause**: 14 robot files used `Run Process    python    -m    cleveragents    ...` with bare `python`. On CI, `session.env["PATH"]` propagation from nox → robot → `Run Process` subprocess was unreliable — the system `/usr/local/bin/python` was used instead of the venv Python, causing `No module named cleveragents`. Files that imported `common.resource` and used `${PYTHON}` (set via `sys.executable` in `Setup Test Environment`) were unaffected.
  - **Fix** (`noxfile.py` + 14 robot files): Noxfile now passes `--variable PYTHON:<venv-python-path>` to the `robot` command, explicitly injecting the venv Python path. All 14 files updated to use `${PYTHON}` instead of bare `python` in `Run Process` calls, with a `${PYTHON}    python` default in `*** Variables ***` for standalone runs. This bypasses PATH entirely.
- **Issue 3 — Hardcoded `/app/src` path in `system_prompt_template_rendering.robot`**:
  - **Root cause**: Three `python -c` inline scripts used `sys.path.insert(0, '/app/src')` to import `cleveragents`. This path is only valid locally (`/app`), not on CI (`/workspace/cleveragents/cleveragents-core`). Since the venv Python (now injected via `${PYTHON}`) already has `cleveragents` installed via `pip install -e .[tests]`, the `sys.path.insert` is unnecessary.
  - **Fix** (`robot/system_prompt_template_rendering.robot`): Removed `import sys; sys.path.insert(0, '/app/src');` from all 3 inline Python code strings.
- **Additional fix**: Added trailing newline to `robot/common.resource` (was missing, defensive fix).
- **CI debug output added** (`noxfile.py`): Comprehensive pre-robot debug block now prints: `sys.executable`, `shutil.which('python')`, full `PATH`, `os.path.exists/isfile/islink/stat` for both `.resource` files, first 100 bytes of each, fixture directory existence checks (`features/fixtures/v2/*`), and Robot Framework version. This will definitively diagnose any remaining resource import issues on the next CI run.
- **Verification**: `nox -s integration_tests` passes locally — 204 tests, 204 passed, 0 failed.

**2026-02-13**: Bugfix - CI `integration_tests` Failures Round 5 (guard cleanup when OpenAI key missing)

- **Observation**: After Round 4, CI still failed mid-suite with a cascade of `Resource file ... does not exist` and `FileNotFoundError` for the venv Python path. The failures started immediately after `Robot.Scientific Paper Basic` was skipped due to missing `OPENAI_API_KEY`.
- **Root cause** (`robot/scientific_paper_basic.robot`): Suite setup begins with `Require OpenAI Key`. When the key is missing, the keyword calls `Skip`, so the rest of setup does not execute. This leaves `${CONTEXT_DIR}` as `${EMPTY}` (it was initialized to empty in the Variables table). In suite teardown, `Remove Directory    ${CONTEXT_DIR}    recursive=True` runs with an empty string, which resolves to the current working directory. On CI, this can delete the repo root (including `.nox` venv and `robot/*.resource` files), causing downstream suites to fail and the venv Python path to disappear.
- **Fix** (`robot/scientific_paper_basic.robot`): Initialize `${CONTEXT_DIR}` to `${TEMPDIR}/paper_basic_contexts` in the Variables table, and guard cleanup with `Run Keyword If    '${CONTEXT_DIR}' != '${EMPTY}'` before removing the directory. This prevents accidental deletion of the workspace when the suite is skipped.
- **Verification**: Not re-run on CI yet; local `nox -s integration_tests` should continue to pass (204/204) with `OPENAI_API_KEY` unset.

**2026-02-13**: Enhancement - Restore parallel Robot runs + silence discovery resource warning

- **Change 1 — Reintroduce `pabot` for `integration_tests`**:
  - **Reason**: Sequential `robot` runs were reliable but slow (~7 minutes on CI). With the teardown bug fixed, we can safely return to parallel execution.
  - **Implementation** (`noxfile.py`): `integration_tests` now runs `pabot` with conservative default parallelism (max 2 processes). Overrides supported via `PABOT_PROCESSES` or `--processes` in nox args. Existing `--variable PYTHON:<venv>` injection remains to ensure `Run Process` uses the venv interpreter.
- **Change 2 — Add `robot/discovery_common.resource` stub**:
  - **Reason**: `robot/core_cli_commands.robot` imports `discovery_common.resource`, which does not exist. This emits a non-fatal warning every run even though discovery-tagged tests are excluded.
  - **Fix**: Add a minimal placeholder resource file to silence the warning.
- **Cleanup**: Remove the temporary CI debug dump in `integration_tests` now that the failure mode has been addressed.

**2026-02-13**: Stage A2b.beta Complete - Plan Model Alignment [Luis]

- Rewrote `src/cleveragents/domain/models/core/plan.py` to align with spec: `processing_state` replaces `action_state`/`state` (per line 1096), `project_links` replaces `project_ids`, added `action_name`, `automation_profile: AutomationProfileRef`, `invariants: list[PlanInvariant]`, `arguments`, actor fields, subplan hierarchy (`SubplanConfig`, `SubplanFailureHandler`, `SubplanStatus`), `as_cli_dict()`, `ProjectLink.validate_alias()`. Enum naming uses `InvariantSource` (not `InvariantScope`). Plan field is `processing_state` with `@property def state` alias.
- Updated 8 source files across domain, application, CLI, and infrastructure layers:
  - `plan_lifecycle_service.py`: `use_action()` accepts `action_name` + `project_links`, all `plan.state =` changed to `plan.processing_state =`.
  - `cli/commands/plan.py`: Uses `plan.project_links` iteration, `action_name=str(action.namespaced_name)`.
  - `infrastructure/database/models.py`: `LifecycleActionModel` and `LifecyclePlanModel` updated for `processing_state`, `project_links` JSON, extra field serialization.
  - `infrastructure/database/repositories.py`: Uses `row.description = action.description`.
  - `domain/models/core/__init__.py`: Exports `InvariantSource`.
  - `cli/commands/action.py`: Added `BusinessRuleViolation` import (from HEAD).
- Updated 12 Behave step definition files and 7 feature files for new field names and removed backwards compatibility shims (no `action phase`, `action state`, `draft`, `make action available`).
- Created `features/plan_model_coverage.feature` (246 lines, 39 scenarios) and `features/steps/plan_model_coverage_steps.py` (489 lines) for comprehensive plan model coverage including `as_cli_dict()`, `ProjectLink` validation, APPLIED auto-correction, alias uniqueness, `is_subplan`/`is_root_plan`/`depth` properties, `SubplanFailureHandler`.
- Created `benchmarks/plan_model_bench.py` (5 suites, 15 benchmarks).
- Updated `docs/reference/plan_model.md` (~270 lines) with phase/state semantics and linkage fields.
- Rebased `feature/m1-plan-model` onto `fd6d41b` (A2b.alpha commit `feat(domain): align action model with spec`). Resolved 30 merge conflicts across source, feature, step, and robot files. Key design decisions during merge:
  1. Plan field name: `processing_state` (per implementation_plan.md line 1096), NOT `state`.
  2. Invariant enum: `InvariantSource` (our naming), NOT `InvariantScope` (HEAD's naming).
  3. Action model (`action.py`): HEAD taken as-is (authoritative for action domain).
  4. No `action_id` on Action domain model — uses `namespaced_name` only.
  5. `make_action_available()`: Kept as no-op in service (from HEAD).
  6. `project_links` used everywhere (NOT `project_ids`).
  7. Feature files: Plan domain features take ours, action domain features take HEAD's.
  8. Step files: Merged — HEAD's `action_name=str(...)` + ours' `project_links=[ProjectLink(...)]`.
- Fixed 6 post-rebase test failures (documented as `Fix -` sub-items under A2b.beta checklist):
  - Missing `@when('I try to make the action available')` step in `plan_lifecycle_service_steps.py`.
  - Missing `@then('the action CLI should make action available')` step in `action_cli_steps.py`.
  - Wrong CLI command (`show` instead of `available`) in `action_cli_steps.py:417`.
  - `state=` instead of `processing_state=` in `plan_lifecycle_commands_coverage_steps.py:58` (Pydantic silently ignored unknown kwarg).
  - Missing required `description` param in `robot/helper_plan_lifecycle_v3.py:672` `create_action()` call.
  - `state=` → `processing_state=` and `project_names` → `project_links` in `robot/helper_db_lifecycle_models.py`.
- Final results: lint 0 errors, typecheck 0 errors, 130 Behave features / 2246 scenarios / 9868 steps all pass, 206 Robot tests all pass, `plan.py` coverage 100% (275 stmts, 56 branches).
- 29 files changed: +2076 / -1314 lines.

**2026-02-06**: CRITICAL ARCHITECTURAL DECISION - Tool-Based Resource Modification
- **REPLACED**: OutputParser/code fence parsing approach
- **WITH**: Tool-based change tracking (modern approach used by Claude Code, Cursor, Aider)
- **Key changes**:
  - LLMs call skills/tools directly (edit_file, write_file, delete_file, etc.)
  - Skills operate on sandbox state directly
  - ChangeSet is built from skill invocation history, NOT by parsing LLM text output
- Added built-in resource skills (now C4.file/C4.search/C4.git): file ops, dir ops, search, git ops
- Added MCP skill adapter (now C7.mcp): connect to external MCP servers
  - Replaced C4 "Multi-File ChangeSet Generation" with "Tool-Based Change Tracking"
  - Added SkillInvocationTracker and ToolCallRouter components
- **Rationale**:
  - No parsing ambiguity (is this code or explanation?)
  - Each operation is explicit, typed, and trackable
  - Supports rollback (replay inverse of recorded changes)
  - Resource-agnostic (works for files, databases, APIs, any resource type)
  - Compatible with MCP standard for external tools
- See `docs/specification.md` sections:
  - "Tool-Based Resource Modification (Modern Architecture)"
  - "Unified Resource Abstraction Layer"
  - "MCP Integration Architecture"

**2026-02-09**: Stage A5.3 + A5.4 Complete - v3 SQLAlchemy Models
- Created `LifecycleActionModel` (table: `actions_v3`) in `src/cleveragents/infrastructure/database/models.py:184`:
  - All columns per spec including action_id (ULID PK), name (unique), namespace, short_name, descriptions, actor refs, inputs_schema (JSON), state, timestamps, tags
  - Indexes on namespace, state, short_name
  - `to_domain()` and `from_domain()` conversion methods
  - Relationship to `LifecyclePlanModel`
- Created `LifecyclePlanModel` (table: `lifecycle_plans`) in `src/cleveragents/infrastructure/database/models.py:327`:
  - All columns per spec including self-referential FKs (parent_plan_id, root_plan_id), action_id FK to actions_v3, phase/state, timestamps (ISO-8601 strings), JSON fields (project_ids, tags, sandbox_refs, execution_log, strategy_context)
  - Indexes on phase, state, parent, root, created_at, action_id
  - `to_domain()` and `from_domain()` conversion methods with `_parse_iso`/`_to_iso` helpers
  - Self-referential parent_plan relationship, relationship to LifecycleActionModel
- Used `__allow_unmapped__ = True` on both models to work with `from __future__ import annotations` and the legacy declarative style
- Updated `src/cleveragents/infrastructure/database/__init__.py` to export both new models

**2026-02-09**: Stage B3.1 + B3.2 + B3.5 + B3.6 + B3.7 + B3.8 Complete - Sandbox Infrastructure (Luis's tasks)
- Created `src/cleveragents/infrastructure/sandbox/` package with 6 modules:
  - `protocol.py` (B3.1 + B3.2): `SandboxStatus` enum with 7 states and transition validation, `SandboxContext` and `CommitResult` frozen dataclasses, `Sandbox` runtime-checkable Protocol, exception hierarchy (`SandboxError`, `SandboxCreationError`, `SandboxCommitError`, `SandboxRollbackError`, `SandboxStateError`)
  - `no_sandbox.py` (B3.5): `NoSandbox` class for non-sandboxable resources (API endpoints, etc.). Implements full Sandbox protocol: `create()` logs warning about immediate changes, `get_path()` returns original path, `commit()` is no-op success, `rollback()` always raises, `cleanup()` idempotent. Path traversal guard included.
  - `factory.py` (B3.6): `SandboxFactory` with `create_sandbox()` method mapping strategy strings to implementations. Currently only `"none"` -> `NoSandbox` is fully implemented; `git_worktree`, `copy_on_write` raise `NotImplementedError` (awaiting Hamza's B3.3/B3.4). Includes `is_supported()` and `get_supported_strategies()` validation helpers with resource type compatibility map.
  - `manager.py` (B3.7): `SandboxManager` with thread-safe (RLock) tracking of active sandboxes per plan. Lazy creation via `get_or_create_sandbox()`, batch operations (`commit_all`, `rollback_all`, `cleanup_all`), abandoned sandbox cleanup, atexit handler for graceful process exit.
  - `merge.py` (B3.8): Three merge strategies: `GitMergeStrategy` (shells out to `git merge-file`, falls back to sequential), `SequentialMergeStrategy` (theirs-wins), `JsonMergeStrategy` (recursive deep-merge with configurable array handling). All implement `MergeStrategy` protocol.
  - `__init__.py`: Exports all public types.
- Design decisions:
  - Factory accepts raw `resource_id`/`original_path`/`sandbox_strategy` strings rather than Resource objects, since Resource model (B1) is not yet implemented by Hamza. This decouples the sandbox layer from the domain model and will be easy to wrap once B1 is complete.
  - `SandboxStatus.assert_transition()` provides a convenient guard that raises `SandboxStateError` on invalid transitions -- used throughout NoSandbox and will be used by GitWorktreeSandbox/FilesystemSandbox.
  - `NoSandbox.rollback()` always raises `SandboxRollbackError` rather than being a no-op, because silently swallowing rollback requests for unsandboxed resources would hide bugs.
- All 6 files pass pyright with 0 errors (39 pre-existing errors in models.py remain).
- Smoke tests pass: NoSandbox lifecycle, status transitions, factory creation, manager get_or_create, merge strategies.
- Remaining Luis sandbox-adjacent tasks: None. Hamza's B3.3 (GitWorktreeSandbox) and B3.4 (FilesystemSandbox) are the remaining sandbox implementations.

**2026-02-09**: Stage A2.1 Already Complete
- `estimation_actor` field already present at `src/cleveragents/domain/models/core/action.py:219`
- `review_actor` field already present at `src/cleveragents/domain/models/core/action.py:211`
- Also has `apply_actor` at line 215 (bonus field from earlier work)
- safety_profile remains DEFERRED to post-30 per plan

**2026-02-09**: Stage SEC1 Complete - Remove eval() Vulnerability (Luis tasks SEC1.1-SEC1.3)
- **SEC1.1 Audit results** (7 matches in `src/cleveragents/`):
  - `reactive/stream_router.py:113` - `exec(code, {}, local_vars)` -- **CRITICAL: arbitrary code execution from config**
  - `reactive/stream_router.py:340` - `eval(fn_str, {"__builtins__": {}})` -- **CRITICAL: eval of config transform strings**
  - `reactive/config_parser.py:64` - `re.compile(...)` -- **SAFE: regex compilation, not Python compile()**
  - `agents/graphs/plan_generation.py:192` - `graph.compile(...)` -- **SAFE: LangGraph compile, not builtin**
  - `agents/graphs/context_analysis.py:135` - `graph.compile(...)` -- **SAFE: LangGraph compile**
  - `agents/graphs/auto_debug.py:74` - `graph.compile(...)` -- **SAFE: LangGraph compile**
  - Result: 2 real vulnerabilities, 4 false positives
- **SEC1.2 Fixes applied** to `src/cleveragents/reactive/stream_router.py` (full removal per NO BACKWARDS COMPATIBILITY rule):
  - `SimpleToolAgent`: Added named operation registry (`_SAFE_OPERATIONS`) with `register_operation()` classmethod. New `operation` param selects from: identity, uppercase, lowercase, strip, extract_content, to_string. **Legacy `code` blocks are fully rejected** -- attempting to use a `code` block raises `StreamRoutingError`. No `exec()` call remains.
  - `ReactiveStreamRouter`: Added named transform registry (`_TRANSFORM_REGISTRY`) with `register_transform()` classmethod. Registry includes: identity, uppercase, lowercase, strip, to_string, extract_content. **Legacy `fn` string expressions that are not in the registry are fully rejected** -- attempting to use an unregistered `fn` string raises `StreamRoutingError`. No `eval()` call remains.
- **SEC1.3**: `config_parser.py` has NO eval/exec/compile vulnerability. The `re.compile()` at line 64 is standard regex compilation. **No changes needed.**
- **Impact on tests**: BDD tests that previously exercised legacy exec/eval code paths have been updated to verify rejection. Tests in `stream_router_agent_tool_coverage.feature`, `stream_router_new_branches.feature`, and `stream_router_uncovered_paths.feature` now assert `StreamRoutingError` for code blocks and unregistered transforms.

**2026-02-10**: Stage SEC1.5 Complete - Security BDD tests (Luis, acting as Rui per plan)
- Created `features/security_eval.feature` with 8 scenarios:
  - Config with code injection attempt is rejected
  - Malicious transform expression does not execute
  - Valid config still works after eval removal (named operation)
  - Named transform from registry works after eval removal
  - Code block with import attempt is rejected
  - Eval expression mimicking registry name is still rejected
  - Custom registered operation works
  - Custom registered transform works
- Created `features/steps/security_eval_steps.py` with step definitions

**2026-02-09**: Stage A5.6 Complete - ActionRepository (Luis tasks A5.6a-A5.6j)
- Created `ActionRepository` class in `src/cleveragents/infrastructure/database/repositories.py`
- Uses session-factory pattern: `__init__(self, session_factory: Callable[[], Session])` -- each method obtains its own session from the factory
- Custom exceptions: `DuplicateActionError` (extends `DatabaseError`), `ActionInUseError` (extends `BusinessRuleViolation`)
- 10 methods implemented:
  - `__init__` + `_session()` helper
  - `create()`: Persists new Action via `LifecycleActionModel.from_domain()`, catches IntegrityError -> DuplicateActionError
  - `get_by_id()`: Query by primary key, returns domain via `to_domain()` or None
  - `get_by_name()`: Query by full namespaced name string, returns domain or None
  - `get_by_namespace()`: Filter by namespace + optional state, order by short_name ASC
  - `get_by_state()`: Filter by state, order by updated_at DESC
  - `update()`: Fetch row, overwrite all mutable columns, auto-set updated_at
  - `list_available()`: Filter state='available' + optional namespace, order by namespace+short_name
  - `delete()`: Check referential integrity against LifecyclePlanModel, raise ActionInUseError if plans exist, else delete
- All public methods decorated with `@database_retry` per ADR-033
- All mutating methods flush (do NOT commit) -- caller/UnitOfWork handles commit
- Uses `Any` type hints for domain Action objects to avoid circular import issues (LifecycleActionModel.to_domain() handles the actual conversion)
- Updated `infrastructure/database/__init__.py` to export ActionRepository, DuplicateActionError, ActionInUseError
- 0 new pyright errors (3 pre-existing sqlalchemy import resolution errors remain)

**2026-02-09**: Stage E1 Complete - Subplan Model (Luis tasks E1.1-E1.5)
- All models added to `src/cleveragents/domain/models/core/plan.py`
- **E1.1a**: `ExecutionMode` enum (SEQUENTIAL, PARALLEL, DEPENDENCY_ORDERED)
- **E1.1b**: `SubplanMergeStrategy` enum (GIT_THREE_WAY, SEQUENTIAL_APPLY, FAIL_ON_CONFLICT, LAST_WINS)
  - Renamed from `MergeStrategy` in spec to `SubplanMergeStrategy` to avoid collision with the infrastructure-level `MergeStrategy` protocol in `infrastructure.sandbox.merge`
- **E1.2a**: `SubplanConfig` Pydantic model with: execution_mode, merge_strategy, max_parallel (1-50, default 5), fail_fast, timeout_per_subplan_seconds, retry_failed, max_retries (0-5, default 2)
- **E1.3a**: Verified `parent_plan_id` and `root_plan_id` already exist in `PlanIdentity` (from A1)
- **E1.3b**: Added `subplan_config: SubplanConfig | None` and `subplan_statuses: list[SubplanStatus]` fields to `Plan`
- **E1.3c**: Added computed properties: `is_subplan` (has parent), `is_root_plan` (no parent or root==self), `depth` (0 for root, -1 placeholder for non-root), `has_subplans` (has statuses)
- **E1.4a**: `SubplanStatus` Pydantic model with: subplan_id, action_name, target_resources, status, timing, results (error, changeset_summary, files_changed), retries (attempt_number, previous_attempts)
- **E1.4b**: `SubplanAttempt` Pydantic model (used Pydantic BaseModel instead of dataclass for consistency)
- **E1.5a**: `SubplanFailureHandler` class with `should_stop_others()` and `should_retry()` methods
- **E1.5b**: `RETRIABLE_FAILURES` and `NON_RETRIABLE_ERRORS` as frozenset constants
- Used Pydantic BaseModel instead of dataclass (per ADR-004) for SubplanAttempt and SubplanStatus
- Added `from __future__ import annotations` to plan.py to support forward references
- 0 pyright errors on plan.py; all smoke tests pass
- E1.6 (Behave tests) is assigned to Rui, skipped

**2026-02-09**: Stage A6 Complete - Automation Levels Foundation (Luis tasks)
- **A6.1**: Added `AutomationLevel` enum to `src/cleveragents/domain/models/core/plan.py:57`:
  - `MANUAL` (default) - user triggers each phase transition
  - `REVIEW_BEFORE_APPLY` - auto strategize+execute, pause before apply
  - `FULL_AUTOMATION` - all phases run without human input
  - Added `automation_level` field to `Plan` model (default MANUAL)
- **A6.2**: Added `default_automation_level` setting to `src/cleveragents/config/settings.py`:
  - String field with default `"manual"`, env var `CLEVERAGENTS_AUTOMATION_LEVEL`
  - Hierarchy: plan-level (explicit) > global default from settings
  - Session-level override deferred (requires session CLI module that doesn't exist yet)
- **A6.3**: Updated `PlanLifecycleService` in `src/cleveragents/application/services/plan_lifecycle_service.py`:
  - `automation_level` parameter on `use_action()` (explicit > global default)
  - `_resolve_automation_level()` reads from settings, falls back to MANUAL
  - `should_auto_progress(plan)` - pure query, checks if plan should auto-advance
  - `auto_progress(plan_id)` - idempotent, advances to next phase if automation permits
  - `complete_strategize()` and `complete_execute()` now call `auto_progress()` automatically
  - `pause_plan(plan_id)` - sets automation to MANUAL to halt auto-progression
  - `resume_plan(plan_id, level)` - restores automation level and triggers immediate auto-progress if ready
  - `set_plan_automation_level(plan_id, level)` - change level for existing non-terminal plan
- **A6.4**: Updated CLI commands in `src/cleveragents/cli/commands/plan.py`:
  - `--automation-level` option on `agents plan use` command (parses and validates)
  - `agents plan set-automation-level <plan_id> <level>` command (calls `set_plan_automation_level`)
  - `config set automation-level` and `session set automation-level` DEFERRED: require new CLI modules (`config.py`, `session.py`) that don't exist yet
- All modified files pass pyright with 0 new errors (only pre-existing structlog import resolution)

**2026-02-10**: Stage A6.5 Complete - Automation Levels BDD Tests (Luis, taking over from Rui)
- Created `features/automation_levels.feature` with 24 Behave scenarios covering:
  - Manual mode: verifies no auto-progression after complete_strategize or complete_execute
  - Review-before-apply mode: auto-executes after strategize, pauses before apply
  - Full automation mode: auto-executes and auto-applies
  - Plan-level override: explicit automation_level on use_action overrides global default
  - Mid-plan level changes: set_plan_automation_level works on non-terminal plans, rejected on terminal
  - should_auto_progress query: verified for all 3 levels in strategize/complete and execute/complete states, plus terminal plans
  - Pause/resume: pause_plan sets level to MANUAL, resume_plan restores level and triggers auto-progress when ready, defaults to REVIEW_BEFORE_APPLY when no level specified
  - Settings resolution: _resolve_automation_level reads from settings, falls back to MANUAL for invalid values
- Created `features/steps/automation_levels_steps.py` with step definitions
- Discovery: pydantic-settings `BaseSettings` does not accept constructor keyword overrides for fields with `validation_alias`; must set attribute after construction. This affects test helpers that need to override `default_automation_level`.
- All 24 new scenarios pass; all 51 existing `plan_lifecycle_service.feature` scenarios continue to pass (no regressions)

---

## Roadmap

### Milestone Overview

| Milestone | Target Date | Description |
|-----------|-------------|-------------|
| **M0: Foundation** | 2026-02-11 (Day 0) | Preserve existing LangGraph infrastructure and tooling baseline. |
| **M1: Local Source Code MVP** | 2026-02-18 (Day 7) | Action YAML -> plan use -> strategize/execute/apply with sandboxed, tool-based change tracking on source code only. |
| **M2: Projects & Resources** | 2026-02-21 (Day 10) | Resource/project registries + CLI, git-checkout + fs-directory support, local sandbox strategies. |
| **M3: Actors, Tools, Skills, Validations** | 2026-02-25 (Day 14) | Actor YAML + compilation, tool/skill/validation registries, built-in file/search/git tools, MCP adapter (local). |
| **M4: Decisions & Invariants** | 2026-03-04 (Day 21) | Decision recording + correction, invariants + automation profiles, plan tree/explain/correct CLI. |
| **M5: Subplans & Parallelism** | 2026-03-08 (Day 25) | Subplan spawning, parallel execution, merge strategies, multi-project support. |
| **M6: Large Project Autonomy** | 2026-03-13 (Day 30) | ACMS v1 (local), deep subplans, autonomous large-project workflows. |
| **M7: Server Stubs (Client Only)** | 2026-03-18 (Day 35) | Client-side server stubs only (no server implementation in this project). |
| **M8: UX & Rendering** | 2026-03-23 (Day 40) | Output rendering formats, UX polish, remaining spec items. |

### Critical Path to 7-Day MVP (Source Code Only)

**WEEK 1 GOAL**: A minimally usable application that can:
1. Register an action from YAML via `agents action create --config <file>` (namespaced name, description, definition_of_done, actors, args, invariants, optional automation profile).
2. Create a project and link a local `git-checkout` resource via `agents project create <name> --resource <resource>` + `agents resource add git-checkout <name> --path <path>` + `agents project link-resource <project> <resource>`.
3. Use the action on one or more projects via `agents plan use [--automation-profile ...] [--invariant ...] [--arg name=value] <action> <project>...` (Strategize phase).
4. Execute in a per-plan sandbox (git_worktree for git-checkout resources) via `agents plan execute <plan_id>` with tool-based change tracking.
5. Apply sandbox changes via `agents plan apply [--yes] <plan_id>` after required validations (if any) pass.

Detailed tasks, ownership, and sequencing live in the Implementation Checklist (Sections 3-6).

### Parallel Workstreams

- Workstream A: Plan lifecycle + action/plan alignment + persistence (Section 3).
- Workstream B: Project/resource registry + CLI + sandbox strategies (Section 4).
- Workstream C: Actor/tool/skill/validation registries + built-in tools (Section 5).
- Workstream D: Tool router + change tracking + apply pipeline (Section 5/6).
- Workstream Q: Quality automation maintenance (Section 0).
- Workstream T: Testing (Brent through 2026-02-27, Rui after).

## Quick Reference for Development

### Tool Commands

```bash
# Development setup
pip install -e .[dev,tests,docs]    # Install with all extras
hatch env create                     # Create virtual environment
hatch shell                          # Activate environment

# Testing
nox                                  # Run all tests
nox -s unit_tests                    # Run Behave tests only
nox -s integration_tests             # Run Robot tests only
nox -s coverage_report               # Check coverage (must be >=97%)

# Code quality
nox -s format                        # Format with Ruff
nox -s lint                          # Lint with Ruff
nox -s typecheck                     # Type check with pyright

# Documentation
nox -s docs                          # Build documentation
```

### Key Files and Their Purpose
- `pyproject.toml` - All project configuration (no setup.py, no requirements.txt)
- `noxfile.py` - All task automation (no Makefile, no scripts/)
- `features/` - Behave unit tests (no tests/ directory)
- `robot/` - Robot integration tests
- `docs/specification.md` - Authoritative specification (source of truth for all architecture)
- `docs/reference/` - Discovery artifacts from Phase 0
- `docs/architecture/decisions/` - ADRs from Phase 1
- `examples/actions/` - Example action YAML config files
- `examples/actors/` - Example actor YAML config files
- `examples/tools/` - Example tool YAML config files
- `examples/skills/` - Example skill YAML config files
- `examples/profiles/` - Example automation profile YAML config files
- `examples/resource-types/` - Example custom resource type YAML config files
- `examples/validations/` - Example validation YAML config files

### CLI Command Groups (spec-defined)
```
agents version | info | diagnostics | init
agents session   create | list | show | delete | export | import | tell
agents project   create | link-resource | unlink-resource | list | show | delete | context (set|show|inspect|simulate)
agents validation add | attach | detach
agents actor     run | add | remove | list | show | context (remove|list|show|export|import|clear)
agents skill     add | remove | list | show | tools
agents tool      add | remove | list | show
agents resource  type (add|remove|list|show) | add | remove | list | show | inspect | tree | link-child | unlink-child
agents plan      list | use | execute | apply | status | cancel | tree | explain | correct | diff | artifacts | prompt | rollback
agents action    create | list | show | archive
agents automation-profile add | remove | list | show
agents config    set | get | list
agents invariant add | list | remove
```

### Environment Variables (CLEVERAGENTS_* only)
```bash
# Core configuration (see spec Configuration section for the full list)
CLEVERAGENTS_HOME=~/.cleveragents
CLEVERAGENTS_LOG_LEVEL=INFO

# Development
CLEVERAGENTS_DEBUG=true
CLEVERAGENTS_TEST_MODE=true
```

## Schedule Adhereance

### 2026-02-12 (Day 2 since kickoff on 2026-02-11)
- Timeline reference: Day 7/M1 = 2026-02-18, Day 10/M2 = 2026-02-21, Day 14/M3 = 2026-02-25, Day 21/M4 = 2026-03-04, Day 25/M5 = 2026-03-08, Day 30/M6 = 2026-03-13, Day 35/M7 = 2026-03-18.
- Snapshot (Team Status | Overall Risk | Pace Notes): **Behind ~8-10 days** | Overall risk = HIGH | Pre-rebaseline pacing, server scope still in queue, Rui unavailable through 2026-02-27.
- Milestone forecast (Target -> ETA | Delta | Risk):
  - M1 (2026-02-18) -> ETA 2026-02-26 | +8d | HIGH
  - M2 (2026-02-21) -> ETA 2026-03-02 | +9d | HIGH
  - M3 (2026-02-25) -> ETA 2026-03-09 | +12d | MED-HIGH
  - M4 (2026-03-04) -> ETA 2026-03-18 | +14d | MED-HIGH
  - M5 (2026-03-08) -> ETA 2026-03-23 | +15d | MEDIUM
  - M6 (2026-03-13) -> ETA 2026-03-31 | +18d | MEDIUM
- Track forecast (Track | Status | ETA | Risk | Blocking):
  - Track A (Plan lifecycle + persistence) | behind ~6-8d | ETA M1+ | HIGH | action/plan alignment + persistence wiring
  - Track B (Projects/resources + sandbox) | not started | ETA M2+ | HIGH | resource registry + sandbox strategies missing
  - Track C (Actors/tools/skills/validations) | not started | ETA M3+ | HIGH | registries + actor YAML missing
  - Track D (Change tracking + apply pipeline) | not started | ETA M1+ | HIGH | tool runtime + change tracking absent
  - Track Q (Quality automation) | ahead | ETA on-time | LOW | gates landed, only maintenance adjustments
  - Track T (Testing) | at risk (Rui out) | ETA slips 4-6d | HIGH | QA load on Brent/Jeff until Rui returns
- Developer forecast (Name | Availability | Load | Risk | Focus):
  - Jeff | available | critical-path overloaded | HIGH | CLI rebaseline + tool router + execute/apply integration
  - Luis | available | high | MEDIUM | persistence + sandbox/change tracking
  - Hamza | available | high | HIGH | resource registry + project linking
  - Aditya | available | medium | MEDIUM | actor YAML + skill config
  - Brent | available | high (QA pickup) | MED-HIGH | tests + nox verification for M1/M2
  - Rui | unavailable 2026-02-13 to 2026-02-27 | zero | HIGH | tasks reassigned; resume after 2026-02-27
  - Mike/Brian | standby | low | LOW | none

### 2026-02-13 (Day 3 since kickoff on 2026-02-11)
- Timeline reference: Day 7/M1 = 2026-02-18, Day 10/M2 = 2026-02-21, Day 14/M3 = 2026-02-25, Day 21/M4 = 2026-03-04, Day 25/M5 = 2026-03-08, Day 30/M6 = 2026-03-13, Day 35/M7 = 2026-03-18.
- Snapshot (Team Status | Overall Risk | Pace Notes): **Behind ~4-6 days** | Overall risk = HIGH | CI rebaseline to Hatch + nox confirmed; M2 resource scope trimmed to git-checkout + fs-mount + fs-directory/file; local-only MVP confirmed.
- Milestone forecast (Target -> ETA | Delta | Risk):
  - M1 (2026-02-18) -> ETA 2026-02-22 | +4d | HIGH
  - M2 (2026-02-21) -> ETA 2026-02-27 | +6d | HIGH
  - M3 (2026-02-25) -> ETA 2026-03-07 | +10d | MED-HIGH
  - M4 (2026-03-04) -> ETA 2026-03-17 | +13d | MED-HIGH
  - M5 (2026-03-08) -> ETA 2026-03-23 | +15d | MEDIUM
  - M6 (2026-03-13) -> ETA 2026-03-31 | +18d | MEDIUM
- Track forecast (Track | Status | ETA | Risk | Blocking):
  - Track A (Plan lifecycle + persistence) | behind ~3-5d | ETA M1+4 | HIGH | action model rebaseline + repo/persistence wiring
  - Track B (Projects/resources + sandbox) | behind ~6-8d | ETA M2+6 | HIGH | resource registry + git_worktree sandbox
  - Track C (Actors/tools/skills/validations) | behind ~7-10d | ETA M3+10 | MED-HIGH | tool/skill/validation registries + actor YAML v3
  - Track D (Change tracking + apply pipeline) | behind ~6-9d | ETA M1+7 | HIGH | tool runtime + changeset + apply gating
  - Track Q (Quality automation) | slightly behind | ETA M1+2 | MED | CI not yet nox/Hatch aligned; coverage gate mismatch
  - Track T (Testing) | at risk (Rui out) | ETA slips 4-6d | HIGH | QA load on Brent + feature owners
- Developer forecast (Name | Availability | Load | Risk | Focus):
  - Jeff | available | critical-path overloaded | HIGH | action model rebaseline, CLI rebaseline, tool runtime + execute/apply integration
  - Luis | available | high | MEDIUM | persistence wiring + sandbox strategy + apply pipeline
  - Hamza | available | high | HIGH | resource registry + project linking + resource type schemas
  - Aditya | available | medium | MEDIUM | actor/skill YAML + hierarchical graph configs
  - Brent | available | high (QA + CI) | MED-HIGH | CLI coverage + nox/CI alignment + integration tests
  - Rui | unavailable 2026-02-13 to 2026-02-27 | zero | HIGH | testing tasks deferred or reassigned
  - Mike/Brian | standby | low | LOW | none

## Implementation Checklist

This comprehensive checklist tracks all implementation tasks for the CleverAgents project. Each phase item includes mandatory **Code**, **Document**, and **Tests** bullets. Only mark the parent complete when every sub-bullet (including any spawned `Fix - ...` remediation tasks) is checked.

**Organization**: This checklist is organized to enable parallel development and achieve a minimally working version as quickly as possible. Workstreams are clearly marked. Dependencies between workstreams are noted at merge points.

Execute all required tests through the appropriate `nox` sessions—never call `behave`, `robot`, or other runners directly. After touching **any** subtask, immediately add discoveries to the Notes section and update task descriptions.

**Commit Ownership Rule**: Each **COMMIT** item has exactly one owner. Every subtask (Code/Docs/Tests/Quality/Commit) must list that same owner in brackets. If a subtask truly requires a different owner, split it into a separate **COMMIT** item under the appropriate parallel group.

**Commit Completion Rule**: A **COMMIT** item is checked only after every subtask is complete, `nox` has passed, coverage is verified >=97%, and `git commit -m "<message>"` (using the commit message in the COMMIT header) has been executed.

**Gitflow Standard (required for every COMMIT item)**:
- Every **COMMIT** block must include explicit Git subtasks with the exact commands for: checkout master, pull master, create/checkout the feature branch, merge `origin/master` into the branch before final tests, `git add .`, and `git commit -m "<message>"`.
- Do **not** list Git command tasks outside a **COMMIT** block; any Git command must be a child of the owning **COMMIT** item.
- Branches follow Gitflow naming: `feature/<section>-<short-name>` (one branch per subtrack unless the subtrack explicitly states otherwise).
- **Merge policy**: merging from a feature branch into `master` is done in **Forgejo via PR only** (no CLI merge to master).
- **Push prohibition**: do **not** include any `git push` commands in checklist items (pushing is handled outside the checklist).
- **PR requirement**: every branch with one or more commits must have a Forgejo PR task with an explicit description; merge happens in the Forgejo UI after CI + review.
- **Branch cleanup**: include local branch deletion commands after PR merge; remote deletion, if desired, should be done in Forgejo UI (no push commands).
- **Compatibility note**: If any existing subtrack still lists CLI merges (e.g., `git checkout master && git merge --no-ff feature/...`), replace with the Forgejo PR step above.

### Updated Team Assignments (by Expertise)

| Developer | Strengths | Assignment Focus | Availability |
|-----------|-----------|------------------|--------------|
| **Jeff** | CTO, fastest developer, expert in everything | Critical path blockers, architecture, complex integrations, decision correction, unblocking others | PRIMARY - Available for all critical work |
| **Aditya** | Domain expert (agents/LLMs), understands hierarchical configs | Actor YAML configurations, hierarchical actor graphs, strategy/execution actors, MCP integration | HIGH - Primary on actor/skill work |
| **Rui** | Fastest developer, new to Python | Testing (Behave/Robot), simple implementations, CLI scaffolding, test fixtures (post-2026-02-27) | UNAVAILABLE 2026-02-13 to 2026-02-27; resumes after |
| **Brent** | Slow but detail-oriented | Days 1-3: Automated quality gates; Days 4-8: QA + test coverage pickup; After Day 8: Validation testing | CONTINUOUS - Independent QA track |
| **Hamza** | RDF expert, Python proficient, no LLM experience | Projects, resources, sandbox, database infrastructure, decision models, context indexing | HIGH - Infrastructure lead |
| **Luis** | Best Python architect, pedantic | Algorithms, state machines, service layer architecture, change tracking, validation pipelines | HIGH - Architecture/service layer |
| **Mike/Brian** | Sysadmins | Deployment only (minimal coding tasks) | LOW - Only deployment tasks |

### Work Assignment Philosophy

**Jeff** should be assigned to:
- Any task that is blocking other developers
- Complex integrations requiring deep architectural understanding
- Decision correction mechanism (critical for 30-day goal)
- Skill execution framework (core to MVP)
- Performance-critical code paths
- Final review of all architectural decisions

### Jeff's Critical Path Task Summary (Day-by-Day)

- Day 1-3: A2b/A4b rebaseline (action/plan domain + CLI), tool registry bootstrap, unblock persistence wiring.
- Day 4-7: Tool router + sandbox integration, plan execute/apply end-to-end, MVP integration gating.
- Day 15-21: Decision correction core algorithm + plan tree/explain/correct CLI integration.
- Day 22-30: Large-project autonomy integration (subplans + ACMS v1 + performance tuning).

**BLOCKING CHAIN**: If Jeff is unavailable, the following chains stall:
- A5.alpha → A5.beta → A5.gamma → Plan persistence (Day 1-2)
- C3.protocol → C3.context → C4.file → Skill execution (Day 3-6)
- D4.revert → D4.append → Decision correction (Day 15-18)

**Aditya** should be assigned to:
- ALL actor configuration YAML files and examples
- Hierarchical actor graph compositions
- Strategy and execution actor templates
- MCP skill adapter implementation
- Skill metadata definitions
- Anything requiring understanding of LLM tool calling patterns

**Luis** should be assigned to:
- State machine implementations (plan lifecycle, phase transitions)
- Repository pattern implementations
- Service layer architecture
- Algorithm design (merge strategies, dependency closure)
- Validation pipeline orchestration
- Should NOT be assigned to simple CRUD or UI work

**Hamza** should be assigned to:
- Database schema design and Alembic migrations
- Resource model and sandbox infrastructure
- Context indexing and RDF graph store integration
- Decision tree persistence layer
- Session management
- Should NOT be assigned to LLM/agent-specific logic

**Rui** should be assigned to:
- ALL Behave test scenarios (write BEFORE implementation)
- Robot integration tests
- Simple CLI scaffolding
- Test fixtures and mocks in `features/`
- Should ALWAYS work in parallel with feature developers
- Availability constraint: Rui is unavailable 2026-02-13 through 2026-02-27; M1/M2 test + CLI tasks must be reassigned during that window.

**Brent** should be assigned to:
- Days 1-3: Setting up comprehensive automated quality gates
- Days 4-8: Selective manual review of high-priority items only
- After Day 8: High-impact validation and edge case testing with Luis
- Continuous: Monitoring automated quality metrics
- Should work INDEPENDENTLY with no blocking dependencies

### Updated Timeline Targets

| Milestone | Day | Goal | Success Criteria |
|-----------|-----|------|------------------|
| **M1: MVP** | Day 7 | Action YAML -> plan use -> execute in sandbox -> apply (source code only) | `agents [--data-dir PATH] [--config-path PATH] action create --config` + `agents [--data-dir PATH] [--config-path PATH] plan use <action> <project>` + `agents [--data-dir PATH] [--config-path PATH] plan execute` + `agents [--data-dir PATH] [--config-path PATH] plan apply` working end-to-end on a git repository with sandboxed execution |
| **M2: Projects & Resources** | Day 10 | Project/resource registry + CLI with git worktree sandbox | `agents project create` + `agents resource add git-checkout` + `agents project link-resource` working with git worktree isolation |
| **M3: Actors, Tools, Skills, Validations** | Day 14 | Actor YAML compiled -> tool/skill/validation registries -> tool-based ChangeSet | Actor YAML parsed -> LangGraph compiled -> Tools executed -> Validations passing -> Applied |
| **M4: Decisions & Invariants** | Day 21 | Decision recording + correction + invariants + automation profiles | `agents plan tree` + `agents plan explain` + `agents plan correct` + `agents invariant add/list` |
| **M5: Subplans & Parallelism** | Day 25 | Hierarchical subplans with parallel execution and merging | Parent plan spawns 5+ subplans -> Parallel execution -> Three-way merge -> Validations pass |
| **M6: Large Projects** | Day 30 | ACMS v1 + large-project autonomy | Port a 500-file codebase using hierarchical subplans + decision correction |
| **Server Connectivity** | Beyond Day 30 | Client interfaces for server communication; server developed independently | Client-to-server API abstractions defined, but local execution only; no server implementation in this project |

### CRITICAL: Server Connectivity Policy

**The CleverAgents client does NOT include server functionality.** The server is a separate project that will be developed independently. This implementation plan covers the **client only**. During Days 1-30, developers should:

1. **Design with Server Connectivity in Mind**: Ensure abstractions support future connection to an external server
2. **Create Client Interface Stubs Only**: Define `ServerClientAPI`, `RemoteExecutionClient`, `AuthenticationClient` interfaces but implement as stubs that raise `NotImplementedError("Server connectivity not yet implemented")`
3. **No Server Implementation Work**: The server is NOT part of this project—do NOT spend time on server implementation, only on client-side connection interfaces
4. **Focus on Core Functionality**: All effort goes to plan lifecycle, actors, skills, sandboxing, decisions, and subplans

The following Section 8 (Server Connectivity) items are explicitly **OUT OF SCOPE** for the 30-day deadline:
- Client-to-server API integration
- WebSocket streaming to server
- Remote project execution via server
- Server authentication flow (client-side)
- Server API client implementation (beyond stubs)

**Rationale**: Getting a minimally usable application working for single-user local operation in 7 days, and large project autonomy in 30 days, is more valuable than incomplete server connectivity. The server will be developed as a separate project.

### Critical Path Analysis

Critical path chains (authoritative tasks live in the checklist sections):

- Chain A (M1): A2b (action/plan rebaseline) → A4b (CLI rebaseline) → A5 (persistence) → A5.legacy (remove legacy plan path) → M1 acceptance.
- Chain B (M2): B1 (resource/project domain + built-ins) → B1/B2 persistence → B3 CLI → B4 sandbox strategies.
- Chain C (M3): C1/C2 actor YAML + loader → tool/skill/validation registries → C5 router/change tracking → C9 execute/apply integration.
- Chain D (M4+): D1/D2 decision model + recording → D4 correction → E1/E2 subplans → ACMS v1 (G-series).

### Parallel Workstream Allocation

Current parallel tracks (see Sections 3-6 for detailed checklists):

- Track A: Plan lifecycle + persistence (Jeff/Luis).
- Track B: Projects/resources + sandbox (Hamza).
- Track C: Actors/tools/skills/validations (Aditya/Jeff).
- Track D: Change tracking + apply/validation integration (Luis/Jeff).
- Track Q: Quality automation (Brent).
- Track T: Testing (Brent through 2026-02-27; Rui after).

### Coordination Checkpoints (Milestone-Level)

- M1: Action YAML → plan use → execute → apply works end-to-end in local mode with sandboxed tool changes and `nox` green.
- M3: Actor YAML compiles, tool/skill/validation registries functional, tool router produces ChangeSet, apply succeeds.
- M4: Decision tree recorded; plan tree/explain/correct works; invariants + automation profiles enforced.
- M6: ACMS v1 + deep subplans supports large-project autonomy (local mode).

### Parallel Workstreams Structure

- Workstream A: Plan lifecycle + persistence (Section 3).
- Workstream B: Projects/resources + sandbox (Section 4).
- Workstream C: Actors/tools/skills/validations (Section 5).
- Workstream D: Change tracking + apply pipeline (Section 6).
- Workstream Q: Quality automation (Section 0).
- Workstream T: Testing (Section T; Brent through 2026-02-27, Rui after).

Merge points and acceptance checks are tracked as checklist items under each milestone section.

---

### Section 0: Quality Automation Setup [WORKSTREAM Q - Brent Lead]

**Target: Days 0-3 (Minimum gates before merges; advanced gates after M1)**

**Parallelization rules**:
- Minimum gating (pre-commit + CI + coverage enforcement) is a merge blocker; it can run in parallel with Week 1 coding but must land before any feature branches merge.
- Advanced automation (complexity metrics, dashboards, extended security scanning) is deferred until after M1 to avoid blocking the MVP.

**Parallel Group Q0-Minimum Gates [Brent - blocks merges]**

- [X] **COMMIT (Owner: Brent | Group: Q0-Minimum | Branch: feature/q0-min-precommit) - Commit message: "feat(qa): add pre-commit baseline hooks"** - done on 2026-02-12 07:21:52 +0000
  - [X] Git [Brent]: `git checkout master`
  - [X] Git [Brent]: `git pull origin master`
  - [X] Git [Brent]: `git checkout -b feature/q0-min-precommit`
  - [X] Git [Brent]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [X] Code [Brent]: Add `pre-commit>=3.6.0` to `pyproject.toml` dev dependencies and ensure it is included in the `dev` extra.
  - [X] Code [Brent]: Create `.pre-commit-config.yaml` pinned to specific hook versions; include `ruff format`, `ruff check`, `pyright`, `check-merge-conflict`, `end-of-file-fixer`, `trailing-whitespace`.
  - [X] Code [Brent]: Add `pyrightconfig.json` validation to pre-commit (hook that fails if config missing or invalid).
  - [X] Code [Brent]: Add/confirm `nox -s lint` session that runs Ruff + pyright using project settings; ensure session exits non-zero on warnings.
  - [X] Code [Brent]: Add/confirm `nox -s format` session for Ruff formatting and align it with pre-commit `ruff format` behavior.
  - [X] Docs [Brent]: Update `CONTRIBUTING.md` with pre-commit install + run steps (no helper scripts).
  - [X] Tests (Behave) [Brent]: Add scenarios in `features/quality_automation.feature` that parse `.pre-commit-config.yaml`, assert required hooks are present, and verify pinned versions.
  - [X] Tests (Robot) [Brent]: Add `robot/quality_automation.robot` that runs `nox -s lint` and asserts zero failures.
  - [X] Tests (ASV) [Brent]: Add `asv/benchmarks/precommit_config_bench.py` to benchmark config parsing and hook list extraction.
  - [X] Quality [Brent]: Run `nox` (all default sessions, including benchmark).
  - [X] Quality [Brent]: Verify coverage >=97% via `nox -s coverage_report`.
  - [X] Git [Brent]: `git add .`
  - [X] Commit [Brent]: `git commit -m "feat(qa): add pre-commit baseline hooks"`.
  - [X] Forgejo PR [Brent]: Open PR from `feature/q0-min-precommit` to `master`, wait for CI + review, merge in UI (no CLI merge)
  - [X] Git [Brent]: `git checkout master`
  - [X] Git [Brent]: `git branch -d feature/q0-min-precommit`

- [ ] **COMMIT (Owner: Brent | Group: Q0-Minimum | Branch: feature/q0-min-ci) - Commit message: "feat(ci): add nox-based PR validation workflow"**
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git pull origin master`
  - [ ] Git [Brent]: `git checkout -b feature/q0-min-ci`
  - [ ] Git [Brent]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [X] Code [Brent]: Update `.forgejo/workflows/ci.yml` to install dependencies via Hatch and run `nox` (unit + integration + typecheck + lint + coverage_report); do not add GitHub workflows. - done on 2026-02-12
  - [X] Code [Brent]: Ensure CI uses Python 3.13, caches pip/Hatch artifacts, and uploads `nox` logs on failure. - done on 2026-02-12
  - [X] Code [Brent]: Fail pipeline if any `nox` session fails or coverage <97% (explicit coverage gate). - done on 2026-02-12
  - [X] Docs [Brent]: Add CI usage notes in `docs/development/ci-cd.md`, including local repro commands and cache notes. - done on 2026-02-12
  - [X] Tests (Behave) [Brent]: Add a scenario that validates the workflow file exists and references required `nox` sessions. - done on 2026-02-12 (features/ci_workflow_validation.feature, 11 scenarios)
  - [X] Tests (Robot) [Brent]: Add a Robot smoke test that runs the same `nox` session matrix locally and asserts zero failures. - done on 2026-02-12 (robot/ci_nox_validation.robot)
  - [X] Tests (ASV) [Brent]: Add `asv/benchmarks/ci_yaml_parse_bench.py` to benchmark workflow parsing and key lookup. - done on 2026-02-12 (benchmarks/ci_yaml_parse_bench.py)
  - [X] Quality [Brent]: Run `nox` (all default sessions, including benchmark). - done on 2026-02-12 (1673 scenarios passed, 0 failed)
  - [X] Quality [Brent]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%. - done on 2026-02-12 (97% total, fail-under=97 passes)
  - [ ] Git [Brent]: `git add .` (only after coverage check passes)
  - [ ] Commit [Brent]: `git commit -m "feat(ci): add nox-based PR validation workflow"`.
  - [ ] Forgejo PR [Brent]: Open PR from `feature/q0-min-ci` to `master`, wait for CI + review, merge in UI (no CLI merge)
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git branch -d feature/q0-min-ci`

- [ ] **COMMIT (Owner: Brent | Group: Q0-Minimum | Branch: feature/q0-min-coverage) - Commit message: "feat(qa): enforce coverage >=97% in CI"**
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git pull origin master`
  - [ ] Git [Brent]: `git checkout -b feature/q0-min-coverage`
  - [ ] Git [Brent]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Brent]: Ensure `nox -s coverage_report` fails below 97% and emits a single-line error message suitable for CI parsing.
  - [ ] Code [Brent]: Update CI summary output (or job annotations) to surface the 97% threshold failure line clearly.
  - [ ] Docs [Brent]: Update `docs/development/testing.md` with the 97% coverage requirement and a sample failure output.
  - [ ] Tests (Behave) [Brent]: Add a scenario that parses coverage config and asserts threshold >=97%.
  - [ ] Tests (Robot) [Brent]: Add a Robot test that runs `nox -s coverage_report` and asserts pass/fail behavior.
  - [ ] Tests (ASV) [Brent]: Add `asv/benchmarks/coverage_report_bench.py` for coverage report runtime baseline.
  - [ ] Quality [Brent]: Run `nox` (all default sessions, including benchmark).
  - [ ] Git [Brent]: `git add .` (only after coverage check passes)
  - [ ] Git [Brent]: `git commit -m "feat(qa): enforce coverage >=97% in CI"` (only after coverage check passes)
  - [ ] Forgejo PR [Brent]: Open PR from `feature/q0-min-coverage` to `master` with description "Enforce 97% coverage via nox coverage_report with explicit CI summary output and updated docs/tests.".
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git branch -d feature/q0-min-coverage`
  - [ ] Quality [Brent]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.

**Parallel Group Q0-Advanced Gates [Brent - AFTER M1]**

- [ ] **COMMIT (Owner: Brent | Group: Q0-Advanced | Branch: feature/q0-adv-security) - Commit message: "feat(qa): align security scans with nox"**
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git pull origin master`
  - [ ] Git [Brent]: `git checkout -b feature/q0-adv-security`
  - [ ] Git [Brent]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Brent]: Ensure `bandit[toml]>=1.7.5` and `semgrep` are in dev dependencies and configured in `pyproject.toml` + `.semgrep.yml`.
  - [ ] Code [Brent]: Ensure pre-commit hooks for Bandit + Semgrep are present with explicit exclude patterns.
  - [ ] Code [Brent]: Align `nox -s security` session to run Bandit + Semgrep with the same configs used in CI.
  - [ ] Docs [Brent]: Document security scan expectations and severity gates in `docs/development/quality-automation.md`.
  - [ ] Tests (Behave) [Brent]: Add a scenario verifying Bandit/Semgrep hooks are declared in `.pre-commit-config.yaml`.
  - [ ] Tests (Robot) [Brent]: Add a Robot test that runs `nox -s security`.
  - [ ] Tests (ASV) [Brent]: Add `asv/benchmarks/security_scan_bench.py` to baseline scan runtime.
  - [ ] Quality [Brent]: Run `nox` (all default sessions, including benchmark).
  - [ ] Git [Brent]: `git add .` (only after coverage check passes)
  - [ ] Git [Brent]: `git commit -m "feat(qa): align security scans with nox"` (only after coverage check passes)
  - [ ] Forgejo PR [Brent]: Open PR from `feature/q0-adv-security` to `master` with description "Align security scanning (Bandit + Semgrep) across pre-commit, nox, and CI with documented severity gates.".
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git branch -d feature/q0-adv-security`
  - [ ] Quality [Brent]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.

- [ ] **COMMIT (Owner: Brent | Group: Q0-Advanced | Branch: feature/q0-adv-complexity) - Commit message: "feat(qa): align complexity checks with nox"**
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git pull origin master`
  - [ ] Git [Brent]: `git checkout -b feature/q0-adv-complexity`
  - [ ] Git [Brent]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Brent]: Ensure `radon>=6.0.1` is in dev dependencies and `nox -s complexity` enforces the agreed thresholds.
  - [ ] Code [Brent]: Align CI to call `nox -s complexity` (non-blocking until M3) and emit a summary line for parsing.
  - [ ] Docs [Brent]: Document complexity thresholds, exception policy, and how to interpret CI output.
  - [ ] Tests (Behave) [Brent]: Add a scenario that asserts radon configuration exists.
  - [ ] Tests (Robot) [Brent]: Add a Robot test that runs `nox -s complexity` on a fixture module.
  - [ ] Tests (ASV) [Brent]: Add `asv/benchmarks/complexity_scan_bench.py` for radon runtime baseline.
  - [ ] Quality [Brent]: Run `nox` (all default sessions, including benchmark).
  - [ ] Git [Brent]: `git add .` (only after coverage check passes)
  - [ ] Git [Brent]: `git commit -m "feat(qa): align complexity checks with nox"` (only after coverage check passes)
  - [ ] Forgejo PR [Brent]: Open PR from `feature/q0-adv-complexity` to `master` with description "Align radon complexity checks with nox + CI outputs and document thresholds.".
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git branch -d feature/q0-adv-complexity`
  - [ ] Quality [Brent]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.

- [ ] **COMMIT (Owner: Brent | Group: Q0-Advanced | Branch: feature/q0-adv-docs) - Commit message: "docs(qa): refresh quality automation guide"**
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git pull origin master`
  - [ ] Git [Brent]: `git checkout -b feature/q0-adv-docs`
  - [ ] Git [Brent]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Docs [Brent]: Refresh `docs/development/quality-automation.md` to reflect nox matrix, CI job names, and coverage >=97% gate.
  - [ ] Docs [Brent]: Link the guide from `README.md` and `CONTRIBUTING.md` (verify links still valid).
  - [ ] Tests (Behave) [Brent]: Add a scenario verifying the guide exists and is linked from README/CONTRIBUTING.
  - [ ] Tests (Robot) [Brent]: Add a Robot doc build smoke test via `nox -s docs`.
  - [ ] Tests (ASV) [Brent]: Add `asv/benchmarks/docs_build_bench.py` for docs build runtime baseline.
  - [ ] Quality [Brent]: Run `nox` (all default sessions, including benchmark).
  - [ ] Git [Brent]: `git add .` (only after coverage check passes)
  - [ ] Git [Brent]: `git commit -m "docs(qa): refresh quality automation guide"` (only after coverage check passes)
  - [ ] Forgejo PR [Brent]: Open PR from `feature/q0-adv-docs` to `master` with description "Refresh quality automation guide with updated nox/CI matrix and coverage gate details.".
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git branch -d feature/q0-adv-docs`
  - [ ] Quality [Brent]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.

---

### Section 1: Completed Foundation (Phases 0-1) [PRESERVED]

- [X] Phase 0: Discovery and Requirements Elaboration
  - [X] Code: Implement Python discovery tooling for CLI, server, data contracts, supporting assets, environment variables, implicit behaviors, and parity matrix generation.
  - [X] Document: Append findings, scripts, and open questions to **Phase 0 Notes** with `file_path:line_number` references.
  - [X] Tests: Run Behave discovery scenarios and Robot smoke suites covering the generated artifacts.

- [X] Phase 1: Architecture Definition
  - [X] Code: Produce ADR stubs, module scaffolding, coding standard configurations, and packaging setup.
    - [X] ADR-001: Python Package Layering and Module Boundaries
    - [X] ADR-002: Asyncio Concurrency Model
    - [X] ADR-003: Dependency Injection Framework
    - [X] ADR-004: Pydantic for Data Validation
    - [X] ADR-005: Error Handling Hierarchy
    - [X] ADR-006: CLEVERAGENTS Environment Variables
    - [X] ADR-007: Repository Pattern for Persistence
    - [X] ADR-008: Provider Plugin Architecture
    - [X] ADR-009: CLI Framework Selection
    - [X] ADR-010: Logging and Observability
  - [X] Document: Update **Phase 1 Notes** with ADR locations, dependency policies, and style guides.
  - [X] Tests: Execute Ruff, pyright, Behave architecture scenarios, and Robot lint pipelines.

---

### Section 2: Preserved Work from Previous Implementation [PRESERVED]

- [X] LangChain/LangGraph Foundation
  - [X] Dependencies installed (langchain, langgraph, langsmith, etc.)
  - [X] ADR-011: LangChain/LangGraph Integration Patterns
  - [X] Base StateGraph classes (BaseAgent, BaseStateGraph)
  - [X] LangChain mock provider (FakeListLLM)

- [X] Core LangGraph Workflows
  - [X] PlanGenerationGraph (load_context -> analyze_requirements -> generate_plan -> validate)
  - [X] ContextAnalysisAgent (5-node workflow for context analysis)
  - [X] AutoDebugGraph (analyze_error -> generate_fix -> validate_fix -> apply_fix)
  - [X] Memory service with EntityMemory
  - [X] CLI streaming integration

- [X] Provider Integration
  - [X] Provider registry
  - [X] OpenAI, Anthropic, Google, OpenRouter adapters
  - [X] LangSmith observability

- [X] Actor System (Stage 7.5)
  - [X] Actor domain model with config hashing
  - [X] Actor persistence (database, repository)
  - [X] Actor registry (built-ins from provider registry)
  - [X] Actor CLI commands (add, update, remove, list, show)
  - [X] Actor-first plan/chat commands (--actor flag)
  - [X] v2 format compatibility for actor configs

---

### Section 3: Plan Lifecycle [WORKSTREAM A - Luis Lead]

**Target: Milestone M1 (+7 days)**

#### Section 3 Notes

- 2026-02-13: A2b.beta (plan model alignment) completed on `feature/m1-plan-model`. Rebased onto A2b.alpha (`fd6d41b`), resolved 30 merge conflicts. Key decisions: `processing_state` field name (not `state`), `InvariantSource` enum (not `InvariantScope`), `project_links` (not `project_ids`), HEAD's `action.py` authoritative. All tests green: 130 Behave features (2246 scenarios), 206 Robot tests, plan.py 100% coverage. See Development Log entry for full details.
- 2026-02-13: New test artifacts created for plan model coverage: `features/plan_model_coverage.feature` (39 scenarios), `features/steps/plan_model_coverage_steps.py` (489 lines), `benchmarks/plan_model_bench.py` (15 benchmarks), `docs/reference/plan_model.md` (~270 lines).
- 2026-02-13: Robot integration test fixes applied post-rebase: `robot/helper_plan_lifecycle_v3.py` (missing `description` param), `robot/helper_db_lifecycle_models.py` (`state=` → `processing_state=`, `project_names` → `project_links`).

**WEEK 1 - CRITICAL PATH**

- [X] **Stage A1: Plan Data Model** (Day 1) - COMPLETED 2026-02-05
  - [X] Code: Create Plan domain model
    - [X] Define `Plan` Pydantic model with fields (plan_id ULID, parent_plan_id, root_plan_id, attempt counter, phase, state, timestamps)
    - [X] Define `PlanPhase` enum (Action, Strategize, Execute, Apply, Applied)
    - [X] Define `PlanState` enum per phase (available, draft, archived, queued, processing, errored, complete, cancelled)
    - [X] Add namespace support to plan naming (`[server:][namespace/]<name>`)
    - [X] Location: `src/cleveragents/domain/models/core/plan.py`
  - [X] Tests: Behave scenarios for plan model validation, phase/state transitions (30 scenarios in `features/plan_model.feature`)

- [X] **Stage A2: Action Model** (Day 1) - COMPLETED 2026-02-05
  - [X] Code: Create Action domain model
    - [X] Define `Action` Pydantic model (name, description, definition_of_done, strategy_actor, execution_actor, inputs_schema, reusable, read_only)
    - [X] Add argument parsing for action parameters (`--arg name:type:required|optional:description`)
    - [X] Location: `src/cleveragents/domain/models/core/action.py`
  - [X] Tests: Behave scenarios for action model validation (22 scenarios in `features/action_model.feature`)
  - [X] **A2.1** [Luis] Extend Action model with additional fields (follow-up):
    - [X] Field `estimation_actor: str | None` - optional actor for cost/risk estimation (already present at `action.py:219`)
    - [X] Field `review_actor: str | None` - optional actor for code review (already present at `action.py:211`)

  **Parallel Group A2b: Action/Plan Spec Rebaseline (M1-critical)**
    **PARALLEL SUBTRACK A2b.alpha [Jeff]**: Action model alignment + YAML-first semantics
    **PARALLEL SUBTRACK A2b.beta [Luis]**: Plan model alignment + action/project linkage
    **PARALLEL SUBTRACK A2b.gamma [Aditya]**: Action YAML schema + examples + loader
    **SEQUENTIAL MERGE NOTE**: A2b.alpha + A2b.beta + A2b.gamma must land before A4b CLI rebaseline.
- [X] **COMMIT (Owner: Jeff | Group: A2b.alpha | Branch: feature/m1-action-model) - Commit message: "feat(domain): align action model with spec"** - done on 2026-02-13
  - [X] Git [Jeff]: `git checkout master`
  - [X] Git [Jeff]: `git pull origin master`
  - [X] Git [Jeff]: `git checkout -b feature/m1-action-model`
  - [X] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [X] Code [Jeff]: Update `src/cleveragents/domain/models/core/action.py` to remove `action_id` and make the namespaced name the unique identifier.
  - [X] Code [Jeff]: Replace `short_description` with required `description`, keep optional `long_description`, and preserve stable serialization order.
  - [X] Code [Jeff]: Align `ActionState` to `available`/`archived` only; default to `available` and remove draft-only behavior.
  - [X] Code [Jeff]: Add optional `automation_profile`, `invariant_actor`, and `invariants` list with trimming, de-dup, and blank rejection.
  - [X] Code [Jeff]: Add optional `review_actor`, `apply_actor`, and `estimation_actor` fields with namespaced validation.
  - [X] Code [Jeff]: Add optional `inputs_schema` (JSON Schema dict) with JSON-serializable validation.
  - [X] Code [Jeff]: Align `ActionArgument` types to spec (`string`, `integer`, `float`, `boolean`, `list`) and update coercion mapping.
  - [X] Code [Jeff]: Add `ActionArgument.from_mapping()` for YAML and `ActionArgument.coerce_value()` for CLI `--arg` inputs.
  - [X] Code [Jeff]: Add `Action.from_config()` and `Action.as_cli_dict()` for YAML-first parsing and stable CLI rendering.
  - [X] Code [Jeff]: Add templating helper for `description`/`definition_of_done` with missing placeholder errors.
  - [X] Docs [Jeff]: Update `docs/reference/action_model.md` for YAML-first fields and state semantics.
  - [X] Tests (Behave) [Jeff]: Add scenarios for namespaced validation, argument coercion, inputs_schema acceptance, and templating errors. (64 scenarios in action_model.feature, 2224 total scenarios passing)
  - [X] Tests (Robot) [Jeff]: Add Robot scenario that loads action YAML and asserts `automation_profile`, `invariants`, and actor refs in CLI output. (Action YAML Config Loading And CLI Dict test in plan_lifecycle_v3.robot)
  - [X] Tests (ASV) [Jeff]: Add `asv/benchmarks/action_model_bench.py` for argument parsing + templating throughput. (6 benchmark suites: TimeArgumentParsing, TimeArgumentCoercion, TimeFromConfig, TimeAsCliDict, TimeTemplateRendering, TimeFromMapping)
  - [X] Quality [Jeff]: Run `nox` (all default sessions, including benchmark). NOTE: unit_tests, typecheck, coverage_report, integration_tests all pass. Benchmark session requires ASV install from git HEAD (uncommitted code not available in ASV env).
  - [X] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. Coverage is 97% overall. Key files: action.py 95%, action CLI 100%, repositories 100%, plan_lifecycle_service 98%.
  - [X] Git [Jeff]: `git add .`
  - [X] Git [Jeff]: `git commit -m "feat(domain): align action model with spec"`
  - [X] Forgejo PR [Jeff]: Open PR from `feature/m1-action-model` to `master` with description "Align action domain model to spec: YAML-first naming, invariants, and argument typing; updates docs + tests."
- [X] **COMMIT (Owner: Luis | Group: A2b.beta | Branch: feature/m1-plan-model) - Commit message: "feat(domain): align plan model with spec"** - done on 2026-02-13
  - [X] Git [Luis]: `git checkout master`
  - [X] Git [Luis]: `git pull origin master`
  - [X] Git [Luis]: `git checkout -b feature/m1-plan-model`
  - [X] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit). NOTE: Rebased onto `fd6d41b` (A2b.alpha) with 30 merge conflict resolutions.
  - [X] Code [Luis]: Remove Action phase from `PlanPhase`; align lifecycle to strategize/execute/apply/applied.
  - [X] Code [Luis]: Replace `action_state` with `processing_state` only; enforce phase/state consistency per spec. Field is `processing_state` with `@property def state` alias.
  - [X] Code [Luis]: Add `action_name` (namespaced) and remove `action_id` usage throughout plan model.
  - [X] Code [Luis]: Replace `project_ids` with `project_links` (name, alias, read_only) and enforce alias uniqueness.
  - [X] Code [Luis]: Add `automation_profile` with provenance tags (plan/action/project/global) and lock after creation. Uses `AutomationProfileRef` model.
  - [X] Code [Luis]: Add `invariants` list with source tags and stable ordering for CLI rendering. Uses `InvariantSource` enum (not `InvariantScope`).
  - [X] Code [Luis]: Add `arguments` map + `arguments_order`, and persist rendered `description`/`definition_of_done` at plan creation.
  - [X] Code [Luis]: Add `changeset_id`, `sandbox_refs`, `validation_summary`, `decision_root_id`, `error_message`, `error_details` placeholders.
  - [X] Code [Luis]: Add `Plan.as_cli_dict()` for stable output and `ProjectLink.validate_alias()` helper.
  - [X] Docs [Luis]: Update `docs/reference/plan_model.md` with phase/state semantics and linkage fields (~270 lines).
  - [X] Tests (Behave) [Luis]: Add scenarios for phase transitions, automation profile provenance, invariant ordering, and project link alias errors. Created `features/plan_model_coverage.feature` (39 scenarios) + `features/steps/plan_model_coverage_steps.py` (489 lines). Total: 75 plan model scenarios (36 in `plan_model.feature` + 39 in `plan_model_coverage.feature`).
  - [X] Tests (Robot) [Luis]: Add Robot scenario asserting `plan status` renders action_name + automation profile provenance (`plan-status-rendering` in `robot/plan_lifecycle_v3.robot`).
  - [X] Tests (ASV) [Luis]: Add `benchmarks/plan_model_bench.py` for plan validation + serialization (5 suites, 15 benchmarks).
  - [X] Quality [Luis]: Run `nox` (all default sessions, including benchmark). NOTE: Benchmark fails because ASV installs from git HEAD; uncommitted code not available in ASV env. All other sessions pass. Results: lint 0 errors, typecheck 0 errors, 130 Behave features / 2246 scenarios / 9868 steps pass, 206 Robot tests pass.
  - [X] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. Coverage: `plan.py` 100% (275 stmts / 56 branches / 0 misses). Overall >=97%.
    - [X] Fix - Missing `@when('I try to make the action available')` step in `features/steps/plan_lifecycle_service_steps.py` (caused `plan_lifecycle_service.feature:73` failure)
    - [X] Fix - Missing `@then('the action CLI should make action available')` step in `features/steps/action_cli_steps.py` (caused `action_cli_uncovered_lines.feature:35` failure)
    - [X] Fix - `action_cli_steps.py:417` used `show` command instead of `available` in CLI invocation (merge artifact)
    - [X] Fix - `plan_lifecycle_commands_coverage_steps.py:58` passed `state=ProcessingState(state)` instead of `processing_state=` to Plan constructor (Pydantic silently ignored unknown kwarg, defaulting to QUEUED). Fixed 4 failing auto-select scenarios.
    - [X] Fix - `robot/helper_plan_lifecycle_v3.py:672` missing required `description` param in `create_action()` call (A2b.alpha renamed `short_description` → `description`)
    - [X] Fix - `robot/helper_db_lifecycle_models.py`: `state=` → `processing_state=` in 3 Plan constructors (lines 135, 213, 239) and `project_names` → `project_links` (line 177)
  - [X] Git [Luis]: `git add .`
  - [X] Git [Luis]: `git commit -m "feat(domain): align plan model with spec"`
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m1-plan-model` to `master`, wait for CI + review, merge in UI (no CLI merge)
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m1-plan-model`
- [ ] **COMMIT (Owner: Aditya | Group: A2b.gamma | Branch: feature/m1-action-schema) - Commit message: "docs(action): add action YAML schema and examples"**
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git pull origin master`
  - [ ] Git [Aditya]: `git checkout -b feature/m1-action-schema`
  - [ ] Git [Aditya]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Docs [Aditya]: Author `docs/schema/action.schema.yaml` with required fields, versioning, and namespaced name validation.
  - [ ] Docs [Aditya]: Add schema blocks for arguments, invariants, automation_profile, inputs_schema, and optional actor refs.
  - [ ] Docs [Aditya]: Add action examples under `examples/actions/` (minimal, invariant-heavy, read-only, estimation-actor, inputs_schema).
  - [ ] Code [Aditya]: Add schema validation helper in `src/cleveragents/action/schema.py` (YAML load + schema validate + env var interpolation).
  - [ ] Code [Aditya]: Normalize YAML keys (snake_case vs camelCase) and emit warnings for legacy aliases.
  - [ ] Code [Aditya]: Normalize invariants list (trim, drop blanks, de-dup) preserving order.
  - [ ] Tests (Behave) [Aditya]: Add scenarios that load each example YAML and assert validation passes; add invalid schema cases.
  - [ ] Tests (Robot) [Aditya]: Add Robot smoke test that parses example YAML files.
  - [ ] Tests (ASV) [Aditya]: Add `asv/benchmarks/action_schema_bench.py` for YAML schema validation throughput.
  - [ ] Quality [Aditya]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Aditya]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Aditya]: `git add .`
  - [ ] Git [Aditya]: `git commit -m "docs(action): add action YAML schema and examples"`
  - [ ] Forgejo PR [Aditya]: Open PR from `feature/m1-action-schema` to `master` with description "Introduce action YAML schema + examples and loader validation; aligns action configs with spec for M1.".
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git branch -d feature/m1-action-schema`

- [X] **Stage A3: Plan State Machine** (Day 1-2) - COMPLETED 2026-02-05
  - [X] Code: Implement plan lifecycle state machine
    - [X] Create `PlanLifecycleService` with phase transition methods
    - [X] Implement `create_action()` - creates plan in Action phase
    - [X] Implement `use_action(action, projects, args)` - transitions to Strategize
    - [X] Implement `execute_plan()` - transitions to Execute
    - [X] Implement `apply_plan()` - transitions to Applied
    - [X] Add validation for phase transitions (only valid transitions allowed)
    - [X] Location: `src/cleveragents/application/services/plan_lifecycle_service.py`
  - [X] Tests: Behave scenarios for all phase transitions, invalid transition errors (29 scenarios in `features/plan_lifecycle_service.feature`)

- [X] **Stage A4: Plan CLI Commands** (Day 2-3) - IN PROGRESS 2026-02-05
  - [X] Code: Implement plan lifecycle CLI
    - [X] `agents [--data-dir PATH] [--config-path PATH] action create --config <file> [<name>] [--strategy-actor <actor>] [--execution-actor <actor>] [--definition-of-done "<text>"] [--arg ...]` (legacy `--name` syntax kept in historical notes)
    - [X] `agents [--data-dir PATH] [--config-path PATH] action list` - list available actions
    - [X] `agents [--data-dir PATH] [--config-path PATH] action show <name>` - show action details
    - [X] `agents [--data-dir PATH] [--config-path PATH] action available <id>` - make action available
    - [X] `agents [--data-dir PATH] [--config-path PATH] action archive <id>` - archive action
    - [X] `agents [--data-dir PATH] [--config-path PATH] plan use <action> <project> [--arg name=value ...]` - create plan from action (legacy `--project` syntax kept in historical notes)
    - [X] `agents [--data-dir PATH] [--config-path PATH] plan execute [plan_id]` - execute current or specified plan
    - [X] `agents [--data-dir PATH] [--config-path PATH] plan apply [plan_id]` - apply executed plan (v3 lifecycle)
    - [X] `agents [--data-dir PATH] [--config-path PATH] plan status [plan_id]` - show plan phase/state
    - [X] `agents [--data-dir PATH] [--config-path PATH] plan list [--phase <phase>] [--state <state>] [--project <project>] [--action <action>]` - list plans with filters
    - [X] `agents [--data-dir PATH] [--config-path PATH] plan cancel <plan_id>` - cancel non-terminal plan
    - [X] Location: `src/cleveragents/cli/commands/action.py`, `src/cleveragents/cli/commands/plan.py`
  - [X] Tests: Behave tests for action CLI (15 scenarios in `features/action_cli.feature`)
  **Parallel Group A4b: CLI Rebaseline + Coverage (M1-critical)**
    **PARALLEL SUBTRACK A4b.alpha [Jeff]**: Action/plan CLI alignment to spec
    **PARALLEL SUBTRACK A4b.beta [Brent]**: Behave + Robot CLI coverage after A4b.alpha
    **SEQUENTIAL MERGE NOTE**: A4b.alpha depends on A2b.alpha/beta/gamma; A4b.beta runs after A4b.alpha to lock outputs.
- [ ] **COMMIT (Owner: Jeff | Group: A4b.alpha | Branch: feature/m1-cli-rebaseline) - Commit message: "feat(cli): rebaseline action and plan lifecycle commands"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m1-cli-rebaseline`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Enforce config-only `agents action create --config` (no legacy flags), add `--update` and `--dry-run`.
  - [ ] Code [Jeff]: Remove `action available`; align `action list/show/archive` to namespaced names only.
  - [ ] Code [Jeff]: Update `plan use` to accept positional projects, parse `--arg name=value`, and apply automation_profile/invariants.
  - [ ] Code [Jeff]: Add actor override flags (`--strategy-actor`, `--execution-actor`, `--estimation-actor`, `--invariant-actor`) and validate actors.
  - [ ] Code [Jeff]: Route CLI to persistence-backed PlanLifecycleService via DI (no in-memory fallback).
  - [ ] Code [Jeff]: Update `plan status/list/execute/apply/cancel` output fields to match spec (action_name, profile, invariants).
  - [ ] Docs [Jeff]: Update `docs/reference/plan_cli.md` and `docs/reference/action_cli.md` with new examples and error cases.
  - [ ] Tests (Behave) [Jeff]: Add CLI scenarios for config-only action create, plan use with invariants/profile, and legacy flag rejection.
  - [ ] Tests (Robot) [Jeff]: Add Robot lifecycle flow for action create -> plan use -> execute -> apply (DB-backed).
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/cli_rebaseline_bench.py` for CLI parsing overhead.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .`
  - [ ] Git [Jeff]: `git commit -m "feat(cli): rebaseline action and plan lifecycle commands"`
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m1-cli-rebaseline` to `master` with description "Rebaseline action/plan CLI to spec (config-only action create, plan use args/invariants, updated outputs) with tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m1-cli-rebaseline`

- [ ] **COMMIT (Owner: Brent | Group: A4b.beta | Branch: feature/m1-cli-tests) - Commit message: "test(cli): cover plan lifecycle commands"**
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git pull origin master`
  - [ ] Git [Brent]: `git checkout -b feature/m1-cli-tests`
  - [ ] Git [Brent]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Tests (Behave) [Brent]: Add coverage for `plan execute`, `plan apply`, `plan status`, `plan list`, `plan cancel` (success + error paths).
  - [ ] Tests (Behave) [Brent]: Add negative cases for invalid automation profile, invalid invariant actor, unknown action.
  - [ ] Tests (Behave) [Brent]: Add output snapshot assertions for table columns and JSON/YAML formats.
  - [ ] Tests (Robot) [Brent]: Add end-to-end Robot suite for lifecycle transitions with multiple projects.
  - [ ] Docs [Brent]: Update `docs/development/testing.md` with new CLI suites and fixtures.
  - [ ] Tests (ASV) [Brent]: Add `asv/benchmarks/plan_cli_smoke_bench.py` for CLI argument parsing overhead.
  - [ ] Quality [Brent]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Brent]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Brent]: `git add .`
  - [ ] Git [Brent]: `git commit -m "test(cli): cover plan lifecycle commands"`
  - [ ] Forgejo PR [Brent]: Open PR from `feature/m1-cli-tests` to `master` with description "Add Behave + Robot CLI coverage for plan lifecycle commands and output formats.".
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git branch -d feature/m1-cli-tests`

**Parallel Group A5: Plan Persistence (M1-critical)**
  **PARALLEL SUBTRACK A5.alpha [Jeff]**: Alembic migrations for action/plan tables
  **PARALLEL SUBTRACK A5.beta [Luis]**: SQLAlchemy models for new tables
  **SEQUENTIAL AFTER alpha+beta [Jeff + Luis]**: Repositories + service integration
  **PARALLEL CONTINUOUS [Brent]**: Persistence tests added inside each commit
  **SEQUENTIAL NOTE**: `action_arguments` migration must land after `actions` migration; A5.legacy should land after A4b CLI alignment + A5.gamma persistence wiring.
  **SEQUENTIAL NOTE**: A5.alpha migrations must match A2b fields; rebase Alembic head after A2b merges before cutting follow-on revisions.
- [X] **COMMIT (Owner: Jeff | Group: A5.alpha | Branch: feature/m1-db-actions) - Commit message: "feat(db): add spec-aligned action and plan tables with migrations, ORM models, and benchmarks"**
  - [X] Git [Jeff]: `git checkout master`
  - [X] Git [Jeff]: `git pull origin master`
  - [X] Git [Jeff]: `git checkout -b feature/m1-db-actions`
  - [X] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [X] Code [Jeff]: Add Alembic migration skeleton with explicit down_revision dependency and naming conventions for indexes/constraints.
  - [X] Code [Jeff]: Create `actions` table with **namespaced_name (PK)**, `namespace`, `name`, and actor refs (strategy/execution/review/apply/estimation/invariant).
  - [X] Code [Jeff]: Add `state` enum column (`available`/`archived`) with default `available` (no draft state).
  - [X] Code [Jeff]: Add description columns (`description`, `long_description`) and `definition_of_done` (rendered text).
  - [X] Code [Jeff]: Add behavioral columns (`automation_profile`, `reusable`, `read_only`, `inputs_schema_json`) and metadata (`tags_json`, `created_by`, timestamps).
  - [X] Code [Jeff]: Add `action_invariants` table with FK to actions by namespaced_name, `invariant_text`, `position`, and created_at.
  - [X] Code [Jeff]: Add unique index on `actions.namespaced_name`, index on `actions.namespace`, and index on `actions.state` for list filters.
  - [X] Code [Jeff]: Ensure downgrade path drops indexes and tables in reverse order.
  - [X] Code [Jeff]: Add Alembic migration for `action_arguments` table with FK to `actions.namespaced_name` and ordered `position` for deterministic argument ordering.
  - [X] Code [Jeff]: Add columns for `name`, `arg_type` (string/integer/float/boolean/list), `requirement`, `description`, `default_value_json`, `min_value`, `max_value`, `validation_pattern`.
  - [X] Code [Jeff]: Add check constraints for numeric min/max ordering and non-empty argument names.
  - [X] Code [Jeff]: Add uniqueness constraint on (action_name, name) and index on (action_name, position).
  - [X] Code [Jeff]: Add Alembic migration for `v3_plans` (replacing `lifecycle_plans`) with ULID PK and identity fields (parent_plan_id, root_plan_id, attempt).
  - [X] Code [Jeff]: Add core plan columns: `namespaced_name`, `namespace`, `description`, `definition_of_done` (rendered).
  - [X] Code [Jeff]: Add lifecycle columns: `phase` enum (strategize/execute/apply/applied), `processing_state` enum (queued/processing/errored/complete/cancelled), and phase timestamps.
  - [X] Code [Jeff]: Add action linkage columns (`action_name` only) and actor refs (strategy/execution/review/apply/estimation/invariant).
  - [X] Code [Jeff]: Add policy/metadata columns (`automation_profile`, `read_only`, `reusable`, `inputs_schema_json`, `created_by`, `tags_json`).
  - [X] Code [Jeff]: Add execution placeholders (`changeset_id`, `sandbox_refs_json`, `validation_summary_json`, `decision_root_id`, `error_message`, `error_details_json`).
  - [X] Code [Jeff]: Add `plan_projects` table with plan_id, project_name (namespaced), alias, read_only flag, and created_at.
  - [X] Code [Jeff]: Add uniqueness constraint on (plan_id, project_name) and index on (project_name) for lookups.
  - [X] Code [Jeff]: Add indexes on `phase`, `processing_state`, and `namespace` for list filtering.
  - [X] Code [Jeff]: Add `plan_arguments` table with plan_id, name, value_json, value_type, and `position` for stable ordering.
  - [X] Code [Jeff]: Add `plan_invariants` table with plan_id, invariant_text, source_scope (plan/action/project/global), optional `position`, and created_at.
  - [X] Code [Jeff]: Add uniqueness constraint on (plan_id, name) for arguments and (plan_id, invariant_text) for invariants.
  - [X] Code [Jeff]: Add index on (plan_id, position) for fast ordered retrieval.
  - [X] Code [Jeff]: Update `LifecycleActionModel` ORM: PK as `namespaced_name`, child relationships for `arguments_rel` and `invariants_rel`, `to_domain()`/`from_domain()` methods.
  - [X] Code [Jeff]: Add `ActionInvariantModel`, `ActionArgumentModel` child ORM models.
  - [X] Code [Jeff]: Update `LifecyclePlanModel` ORM: maps to `v3_plans`, `action_name` FK, `processing_state`, child relationships for `project_links_rel`, `arguments_rel`, `invariants_rel`, `to_domain()`/`from_domain()` methods.
  - [X] Code [Jeff]: Add `PlanProjectModel`, `PlanArgumentModel`, `PlanInvariantModel` child ORM models.
  - [X] Code [Jeff]: Update `ActionRepository` to use `namespaced_name` identity, update child table management in `update()`.
  - [X] Code [Jeff]: Export all new child models from `infrastructure.database.__init__`.
  - [X] Docs [Jeff]: Create `docs/reference/database_schema.md` with column-level details, constraints, indexes, FK relationships, and ER diagram for all 7 new tables.
  - [X] Docs [Jeff]: Document argument storage, JSON serialization rules, and invariant source scopes.
  - [X] Tests (Behave) [Jeff]: Update migration scenarios and model coverage steps for new column names and child table patterns.
  - [X] Tests (Robot) [Jeff]: Update Robot integration tests for new schema (namespaced_name lookups, child table inserts).
  - [X] Tests (ASV) [Jeff]: Add `benchmarks/db_migration_actions_bench.py` for action ORM round-trip baseline.
  - [X] Tests (ASV) [Jeff]: Add `benchmarks/db_migration_action_args_bench.py` for argument serialization baseline.
  - [X] Tests (ASV) [Jeff]: Add `benchmarks/db_migration_plans_bench.py` for plan ORM round-trip baseline.
  - [X] Quality [Jeff]: Run `nox -s typecheck` -- 0 errors, 0 warnings.
  - [X] Quality [Jeff]: Run `nox -s lint` -- All checks passed.
  - [X] Quality [Jeff]: Run `nox -s unit_tests` -- 130 features passed, 0 failed; 2233 scenarios passed, 0 failed.
  - [X] Quality [Jeff]: Run `nox -s integration_tests -- --include database` -- 8 tests, 8 passed.
  - [X] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report` -- TOTAL 97%.
  - [X] Git [Jeff]: `git add .`
  - [X] Git [Jeff]: `git commit -m "feat(db): add spec-aligned action and plan tables with migrations, ORM models, and benchmarks"`
  - [X] Forgejo PR [Jeff]: Open PR from `feature/m1-db-actions` to `master` with description "Add spec-aligned action/plan tables (actions, action_invariants, action_arguments, v3_plans, plan_projects, plan_arguments, plan_invariants) with Alembic migrations, updated ORM models, repository changes, schema docs, and ASV benchmarks.".
  - [X] Git [Jeff]: `git checkout master`
  - [X] Git [Jeff]: `git branch -d feature/m1-db-actions`
- [ ] **COMMIT (Owner: Luis | Group: A5.beta | Branch: feature/m1-orm-models) - Commit message: "feat(models): add action and lifecycle plan ORM models"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m1-orm-models`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [X] Code [Luis]: Add SQLAlchemy base model mixins for ULID PKs, timestamps, and JSON columns (reused by plan/resource models).
  - [X] Code [Luis]: Implement `ActionModel` with PK `namespaced_name`, namespace/name, description/long_description, actor refs, DoD, automation_profile, invariant_actor, state (available/archived), inputs_schema_json, tags_json, created_by.
  - [X] Code [Luis]: Implement `ActionInvariantModel` with FK to actions by namespaced_name, invariant_text, position, and created_at.
  - [X] Code [Luis]: Implement `ActionArgumentModel` with FK to actions by namespaced_name, name, arg_type, requirement, defaults/min/max/regex, position, and constraints.
  - [X] Code [Luis]: Implement `LifecyclePlanModel` with identity fields, phase/processing enums (strategize/execute/apply), action_name linkage, rendered DoD/description, policy metadata, and execution placeholders.
  - [X] Code [Luis]: Implement `PlanProjectLinkModel` with plan_id, project_name, alias, read_only, and created_at, plus uniqueness constraint.
  - [X] Code [Luis]: Implement `PlanArgumentModel` and `PlanInvariantModel` with ordered `position` fields and source_scope for invariants.
  - [X] Code [Luis]: Define ORM relationships with ordering (`order_by=position`) and cascade rules for argument/invariant collections.
  - [X] Code [Luis]: Implement `to_domain()` and `from_domain()` mappers with enum conversion and timestamp normalization.
  - [X] Code [Luis]: Add repository-facing helpers for filtering by namespace/phase/state and eager-load action arguments + plan links.
  - [ ] Docs [Luis]: Update ORM mapping notes in `docs/reference/database_schema.md` with model field mapping table.
  - [ ] Tests (Behave) [Luis]: Add scenarios for ORM round-trip serialization, enum conversions, and ordered argument persistence.
  - [ ] Tests (Robot) [Luis]: Add Robot test that loads a plan and asserts field mapping correctness.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/orm_mapping_bench.py` for Action/Plan mapping throughput.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .`
  - [ ] Git [Luis]: `git commit -m "feat(models): add action and lifecycle plan ORM models"`
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m1-orm-models` to `master` with description "Add/align ORM models for action + plan persistence to match spec fields and ordering.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m1-orm-models`
- [ ] **COMMIT (Owner: Luis | Group: A5.beta2 | Branch: feature/m1-orm-models-refactor) - Commit message: "refactor(models): align action/plan ORM with spec IDs"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m1-orm-models-refactor`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Update `LifecycleActionModel` to use namespaced name as the primary key (remove `action_id` ULID) and align column names with spec (`description`, `long_description`, `automation_profile`, `invariant_actor`, `inputs_schema_json`, `tags_json`).
  - [ ] Code [Luis]: Update `LifecyclePlanModel` to reference `action_name` (namespaced) instead of `action_id`; replace `project_ids` with serialized `project_links` per spec.
  - [ ] Code [Luis]: Add/align plan metadata fields (`automation_profile`, `invariants`, `arguments` + `arguments_order`, `validation_summary`, `decision_root_id`, `error_details_json`) and remove legacy fields no longer in spec.
  - [ ] Code [Luis]: Align `processing_state` enum values to spec and remove legacy `state` values that no longer apply.
  - [ ] Code [Luis]: Update ORM relationships to use namespaced names and ensure ordering for arguments/invariants/project links.
  - [ ] Code [Luis]: Update `to_domain()`/`from_domain()` mappers for actions/plans to reflect new identifiers and serialized fields.
  - [ ] Docs [Luis]: Update ORM mapping notes in `docs/reference/database_schema.md` to reflect the new identifiers and column mappings.
  - [ ] Tests (Behave) [Luis]: Add scenarios verifying ORM round-trip with namespaced IDs and project links.
  - [ ] Tests (Robot) [Luis]: Add Robot test that loads an action + plan and asserts new field mappings.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/orm_refactor_bench.py` for mapping throughput baseline.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Git [Luis]: `git commit -m "refactor(models): align action/plan ORM with spec IDs"` (only after coverage check passes)
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m1-orm-models-refactor` to `master` with description "Align ORM models with spec identifiers (namespaced action IDs, plan action_name linkage, project_links serialization) and update mappings/tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m1-orm-models-refactor`
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
- [ ] **COMMIT (Owner: Jeff | Group: A5.gamma | Branch: feature/m1-repositories) - Commit message: "feat(repo): add action and lifecycle plan repositories"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m1-repositories`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Define repository interfaces in `src/cleveragents/domain/repositories/` for ActionRepository and PlanRepository (methods + expected errors).
  - [X] Code [Luis]: Implement ActionRepository CRUD keyed by namespaced_name with deterministic ordering (created_at) and filters (namespace, state, automation_profile). (completed by Luis)
  - [ ] Code [Jeff]: Implement ActionRepository persistence for arguments + invariants with ordered `position` preservation.
  - [ ] Code [Jeff]: Implement PlanRepository CRUD with filters (phase/state/project_name/action_name) and lookup by plan_id/ namespaced_name.
  - [ ] Code [Jeff]: Implement PlanRepository persistence for plan_projects, plan_arguments, plan_invariants with ordered retrieval.
  - [ ] Code [Jeff]: Add retry decorator to repositories; retry only on `OperationalError`, never on `IntegrityError`.
  - [ ] Code [Jeff]: Add pagination parameters (`limit`, `offset`) with default ordering for list queries.
  - [ ] Docs [Jeff]: Document repository interfaces, error types, and pagination guidance.
  - [ ] Tests (Behave) [Jeff]: Add scenarios for repository create/get/list/update/delete guardrails, including action argument round-trips.
  - [ ] Tests (Robot) [Jeff]: Add Robot test that exercises repository through service layer.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/repository_query_bench.py` for list + filter performance.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .`
  - [ ] Git [Jeff]: `git commit -m "feat(repo): add action and lifecycle plan repositories"`
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m1-repositories` to `master` with description "Add action + plan repositories and repository interfaces with ordered persistence and filters.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m1-repositories`
- [ ] **COMMIT (Owner: Luis | Group: A5.gamma | Branch: feature/m1-lifecycle-service) - Commit message: "feat(service): persist plan lifecycle via repositories"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m1-lifecycle-service`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Update `PlanLifecycleService` to use repositories instead of in-memory dicts.
  - [ ] Code [Luis]: Replace in-memory action/plan maps with repository lookups in `get_action`, `get_plan`, and list helpers.
  - [ ] Code [Luis]: Persist action creation keyed by namespaced_name with arguments/invariants and enforce uniqueness.
  - [ ] Code [Luis]: Persist plan creation with project link metadata (alias/read_only) and store plan_arguments/plan_invariants in the same transaction.
  - [ ] Code [Luis]: Update phase transitions to strategize/execute/apply only and treat apply+complete as terminal applied.
  - [ ] Code [Luis]: Wrap transitions in UnitOfWork transactions and map DB errors to domain errors (duplicate names, missing action).
  - [ ] Code [Luis]: Add optimistic guards for phase transitions (validate phase/state before update; reload on conflict).
  - [ ] Docs [Luis]: Update service docs to reflect persistence and remove in-memory notes.
  - [ ] Tests (Behave) [Luis]: Add scenarios for persisted lifecycle transitions and error handling (duplicate names, invalid transitions).
  - [ ] Tests (Robot) [Luis]: Add end-to-end test that restarts the app and re-reads plan state.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/plan_lifecycle_service_bench.py` for persistence operations.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .`
  - [ ] Git [Luis]: `git commit -m "feat(service): persist plan lifecycle via repositories"`
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m1-lifecycle-service` to `master` with description "Persist plan lifecycle via repositories + UoW; remove in-memory lifecycle storage.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m1-lifecycle-service`

- [ ] **COMMIT (Owner: Luis | Group: A5.gamma | Branch: feature/m1-lifecycle-service) - Commit message: "feat(di): wire lifecycle repos and services"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m1-lifecycle-service`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Register ActionRepository + PlanRepository in `application/container.py` and UnitOfWork.
  - [ ] Code [Luis]: Register PlanLifecycleService with repository dependencies and settings in container.
  - [ ] Code [Luis]: Inject lifecycle service into CLI commands; remove direct service instantiation in `action.py` and `plan.py`.
  - [ ] Code [Luis]: Add container wiring tests to ensure singleton lifetimes are correct and repositories share UoW session.
  - [ ] Docs [Luis]: Update DI wiring notes in `docs/architecture/decisions/adr-003.md`.
  - [ ] Tests (Behave) [Luis]: Add scenarios that use container wiring for lifecycle commands.
  - [ ] Tests (Robot) [Luis]: Add Robot smoke test verifying CLI uses persisted service.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/di_container_bench.py` for container resolution overhead.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .`
  - [ ] Git [Luis]: `git commit -m "feat(di): wire lifecycle repos and services"`
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m1-lifecycle-service` to `master` with description "Wire lifecycle repositories + services in DI container and CLI entrypoints.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m1-lifecycle-service`

- [ ] **COMMIT (Owner: Brent | Group: A5.tests | Branch: feature/m1-persistence-tests) - Commit message: "test(persistence): add plan/action persistence suites"**
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git pull origin master`
  - [ ] Git [Brent]: `git checkout -b feature/m1-persistence-tests`
  - [ ] Git [Brent]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Tests (Behave) [Brent]: Add plan persistence scenarios (create, update phase/state, list filters, plan tree, concurrency).
  - [ ] Tests (Behave) [Brent]: Add action persistence scenarios (create, list available, archive guard, action arguments persisted).
  - [ ] Tests (Behave) [Brent]: Add cross-restart scenarios verifying plan status persists across process restarts.
  - [ ] Tests (Behave) [Brent]: Add migration-boundary scenarios to ensure new tables are read after upgrade.
  - [ ] Tests (Robot) [Brent]: Add plan persistence E2E (full lifecycle, restart persistence, concurrent CLI access).
  - [ ] Docs [Brent]: Update `docs/development/testing.md` with persistence suites, fixtures, and `nox` commands.
  - [ ] Tests (ASV) [Brent]: Add `asv/benchmarks/persistence_suites_bench.py` for DB read/write baselines.
  - [ ] Quality [Brent]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Brent]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Brent]: `git add .`
  - [ ] Git [Brent]: `git commit -m "test(persistence): add plan/action persistence suites"`
  - [ ] Forgejo PR [Brent]: Open PR from `feature/m1-persistence-tests` to `master` with description "Add Behave/Robot persistence coverage for action + plan repositories and migrations.".
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git branch -d feature/m1-persistence-tests`

- [X] **Stage A6: Automation Levels Foundation** (Day 4-5) **[Luis]**
  - [X] Code: Implement basic automation level support
    - [X] **A6.1** [Luis] Add `AutomationLevel` enum to `src/cleveragents/domain/models/core/plan.py`:
      - [X] Value `MANUAL` - user triggers each phase transition
      - [X] Value `REVIEW_BEFORE_APPLY` - auto strategize+execute, pause before apply
      - [X] Value `FULL_AUTOMATION` - all phases automatic
    - [X] **A6.2** [Luis] Add automation level configuration to `src/cleveragents/config/settings.py`:
      - [X] Add `default_automation_level: AutomationLevel` setting
      - [X] Add `CLEVERAGENTS_AUTOMATION_LEVEL` environment variable
      - [X] Implement hierarchy: plan-level > session-level > global-level
    - [X] **A6.3** [Luis] Update `PlanLifecycleService` to respect automation levels:
      - [X] Add `automation_level` parameter to `use_action()` method
      - [X] If automation allows, automatically call `execute_plan()` after strategize completes
      - [X] If full automation, automatically call `apply_plan()` after execute completes
      - [X] Add pause/resume capability for review-before-apply mode
    - [X] **A6.4** [Luis] Update CLI commands to support automation levels:
      - [X] Add `--automation-level` flag to `agents [--data-dir PATH] [--config-path PATH] plan use` command
      - [ ] Add `agents [--data-dir PATH] [--config-path PATH] config set automation-level <level>` command
      - [X] Add `agents [--data-dir PATH] [--config-path PATH] plan set-automation-level <plan_id> <level>` command:
        - [X] Can change automation level for existing plan
        - [X] Only affects future phase transitions
        - [X] Subplans created after change use new level
      - [ ] Add `agents [--data-dir PATH] [--config-path PATH] session set automation-level <level>` command:
        - [ ] Set session-level automation (overrides global)
        - [ ] Persists for current session only
  - [X] Tests: Automation level tests
    - [X] **A6.5** [Luis] Write Behave scenarios in `features/automation_levels.feature` (24 scenarios):
      - [X] Scenario: Manual mode requires explicit execute command
      - [X] Scenario: Manual mode requires explicit apply command
      - [X] Scenario: Review-before-apply auto-executes after strategize completes
      - [X] Scenario: Review-before-apply pauses at apply
      - [X] Scenario: Full automation auto-executes after strategize completes
      - [X] Scenario: Full automation auto-applies after execute completes
      - [X] Scenario: Plan-level automation overrides global setting
      - [X] Scenario: Use action with explicit automation level
      - [X] Scenario: Use action without explicit automation level uses global default
      - [X] Scenario: Change automation level mid-plan works correctly
      - [X] Scenario: Cannot change automation level on terminal plan
      - [X] Scenario: should_auto_progress returns false for manual plan in strategize complete
      - [X] Scenario: should_auto_progress returns true for review-before-apply plan in strategize complete
      - [X] Scenario: should_auto_progress returns false for review-before-apply plan in execute complete
      - [X] Scenario: should_auto_progress returns true for full-automation plan in execute complete
      - [X] Scenario: should_auto_progress returns false for terminal plan
      - [X] Scenario: Pause plan sets automation to manual
      - [X] Scenario: Cannot pause terminal plan
      - [X] Scenario: Resume plan restores automation level
      - [X] Scenario: Resume plan without explicit level defaults to review-before-apply
      - [X] Scenario: Cannot resume terminal plan
      - [X] Scenario: Resume plan triggers auto-progress when ready
      - [X] Scenario: Resolve automation level from settings
      - [X] Scenario: Invalid automation level in settings falls back to manual

**Parallel Group A5.legacy: Remove legacy plan build/apply path (M1-critical)**
- [ ] **COMMIT (Owner: Jeff | Group: A5.legacy | Branch: feature/m1-remove-legacy-plan) - Commit message: "refactor(plan): remove legacy plan service and CLI"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m1-remove-legacy-plan`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Remove `PlanService` usage from CLI (`plan tell/build/apply/new/current/list/cd/continue`) and delete command handlers from `src/cleveragents/cli/commands/plan.py`.
  - [ ] Code [Jeff]: Remove or quarantine legacy `plan_service.py`, `plan_legacy.py`, and legacy CLI helpers; add explicit NotImplementedError where needed.
  - [ ] Code [Jeff]: Remove `PlanService` wiring from `src/cleveragents/application/container.py` and any references from `cli/commands/auto_debug.py`.
  - [ ] Code [Jeff]: Rename `plan lifecycle-apply` to `plan apply` and update command wiring once legacy apply is removed.
  - [ ] Code [Jeff]: Remove or archive legacy `PlanModel`/`PlanStatus` DB tables if unused by v3; document migration path.
  - [ ] Docs [Jeff]: Update CLI docs to list only v3 lifecycle commands and new `plan use/execute/apply` flows.
  - [ ] Tests (Behave) [Jeff]: Remove/replace legacy scenarios with v3 equivalents and adjust coverage expectations.
  - [ ] Tests (Robot) [Jeff]: Remove legacy robot suites and add v3 replacements where needed.
  - [ ] Tests (ASV) [Jeff]: Update asv suite to remove legacy plan benchmarks and add v3 lifecycle baseline benchmark.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .`
  - [ ] Git [Jeff]: `git commit -m "refactor(plan): remove legacy plan service and CLI"`
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m1-remove-legacy-plan` to `master` with description "Remove legacy plan build/apply CLI and services after v3 lifecycle is fully wired.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m1-remove-legacy-plan`

**Parallel Group A6: Automation Profiles Foundation [Jeff + Luis]** (M4-critical; depends on A5 persistence)
  **PARALLEL SUBTRACK A6.core [Jeff]**: Profile model + built-ins + schema
  **PARALLEL SUBTRACK A6.service [Luis]**: Profile resolution + precedence
  **PARALLEL SUBTRACK A6.cli [Jeff]**: CLI commands for profiles
  **SEQUENTIAL MERGE NOTE**: A6.core must land before A6.service/cli; A6.service must land before gating integration in Section 6.
- [ ] **COMMIT (Owner: Jeff | Group: A6.core | Branch: feature/m4-automation-profiles-core) - Commit message: "feat(domain): add automation profile model and built-ins"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m4-automation-profiles-core`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [X] Code [Luis]: Add `AutomationLevel` enum (manual/review/full) in plan domain model. (completed by Luis as foundation)
  - [X] Code [Luis]: Add config setting + env var (`CLEVERAGENTS_AUTOMATION_LEVEL`) with precedence plan > session > global. (completed by Luis)
  - [ ] Code [Jeff]: Add `AutomationProfile` model with **10 confidence thresholds + 3 boolean safety flags** per spec; validate 0.0–1.0 ranges and boolean types.
  - [ ] Code [Jeff]: Add built-in profiles (`manual`, `review`, `supervised`, `cautious`, `trusted`, `auto`, `ci`, `full-auto`) with exact threshold values per spec and stable IDs.
  - [ ] Code [Jeff]: Add YAML schema for automation profiles under `docs/schema/automation_profile.schema.yaml` and loader helper.
  - [ ] Docs [Jeff]: Add `docs/reference/automation_profiles.md` describing built-ins and threshold semantics.
  - [X] Tests (Behave) [Luis]: Add scenarios for default and override precedence (24 scenarios in `features/automation_levels.feature`).
  - [ ] Tests (Behave) [Jeff]: Add scenarios for profile validation and built-in defaults.
  - [ ] Tests (Robot) [Jeff]: Add Robot test that loads each built-in profile and prints summary.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/automation_profile_bench.py` for profile validation.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .`
  - [ ] Git [Jeff]: `git commit -m "feat(domain): add automation profile model and built-ins"`
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m4-automation-profiles-core` to `master` with description "Add automation profile domain model, built-ins, schema, and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m4-automation-profiles-core`
- [ ] **COMMIT (Owner: Luis | Group: A6.service | Branch: feature/m4-automation-profiles-service) - Commit message: "feat(service): resolve automation profiles with precedence"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m4-automation-profiles-service`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [X] Code [Luis]: Add `automation_level` handling to `use_action()` and auto transition logic for execute/apply. (completed by Luis)
  - [X] Code [Luis]: Add pause/resume behavior for review-before-apply mode. (completed by Luis)
  - [ ] Code [Luis]: Add `AutomationProfileService` to resolve profiles with precedence (plan > action > project > global).
  - [ ] Code [Luis]: Add persistence table `automation_profiles` (namespaced name PK) and repository with list/show/update.
  - [ ] Code [Luis]: Add config key `core.automation_profile` and env var override for global default.
  - [ ] Docs [Luis]: Update `docs/reference/config.md` with automation profile defaults and override behavior.
  - [ ] Tests (Behave) [Luis]: Add scenarios for precedence resolution and missing profile errors.
  - [ ] Tests (Robot) [Luis]: Add Robot config smoke test for global profile override.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/automation_profile_resolution_bench.py` for resolution latency.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .`
  - [ ] Git [Luis]: `git commit -m "feat(service): resolve automation profiles with precedence"`
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m4-automation-profiles-service` to `master` with description "Add automation profile resolution service, persistence, and config precedence with tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m4-automation-profiles-service`
- [ ] **COMMIT (Owner: Jeff | Group: A6.cli | Branch: feature/m4-automation-profiles-cli) - Commit message: "feat(cli): add automation-profile commands"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m4-automation-profiles-cli`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [X] Code [Luis]: Add `--automation-level` to `plan use` and add `plan set-automation-level` command. (completed by Luis)
  - [ ] Code [Jeff]: Implement `agents automation-profile add/remove/list/show` commands with YAML config input and schema version guard.
  - [ ] Code [Jeff]: Add `--update` behavior with clear conflict errors; preserve original created_at on update.
  - [ ] Code [Jeff]: Ensure `automation-profile list` supports `--namespace` and regex filters with deterministic ordering.
  - [ ] Code [Jeff]: Add `--format json/yaml` output for `list/show` and keep field names aligned with schema.
  - [ ] Code [Jeff]: Add `--automation-profile` to `plan use` and surface profile name + thresholds in `plan status` output.
  - [ ] Code [Luis]: Add `config set automation-level` and `session set automation-level` commands.
  - [ ] Docs [Jeff]: Update CLI reference with automation-profile command examples, built-in profile list, and error cases.
  - [ ] Tests (Behave) [Jeff]: Add CLI scenarios for profile add/list/show/remove/update and invalid name/threshold validation errors.
  - [ ] Tests (Behave) [Jeff]: Add output snapshot assertions for `automation-profile list --format json` and `show --format yaml`.
  - [ ] Tests (Robot) [Jeff]: Add Robot CLI tests for automation-profile commands (add/show/remove + format outputs).
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/automation_profile_cli_bench.py` for CLI parsing.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .`
  - [ ] Git [Jeff]: `git commit -m "feat(cli): add automation-profile commands"`
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m4-automation-profiles-cli` to `master` with description "Add automation-profile CLI commands, output formats, and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m4-automation-profiles-cli`

**Parallel Group A4c: Plan Diagnostics & Artifacts [Jeff]** (post-M1; depends on C5.diff + C9.execute)
  **SEQUENTIAL NOTE**: A4c should not start until ChangeSet diff artifacts are stable (C5.diff) and plan execution writes plan metadata (C9.execute/C9.apply).
- [ ] **COMMIT (Owner: Jeff | Group: A4c.plan-diff | Branch: feature/m4-plan-diagnostics) - Commit message: "feat(cli): add plan diff command"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m4-plan-diagnostics`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Add `agents plan diff <plan_id>` and `agents plan diff --correction <correction_id>` CLI wiring with rich/plain/json/yaml output.
  - [ ] Code [Jeff]: Load ChangeSet + DiffBuilder artifacts for the plan or correction attempt, fail with explicit error when missing.
  - [ ] Code [Jeff]: Normalize diff output grouping by resource with stable ordering for snapshot tests.
  - [ ] Docs [Jeff]: Update CLI reference with diff examples and `--correction` usage.
  - [ ] Tests (Behave) [Jeff]: Add scenarios for diff output (single resource, multi-resource, missing diff, correction diff).
  - [ ] Tests (Robot) [Jeff]: Add Robot CLI test running `plan diff` after an execute/apply dry run.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/plan_diff_cli_bench.py` for diff rendering overhead.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .`
  - [ ] Git [Jeff]: `git commit -m "feat(cli): add plan diff command"`
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m4-plan-diagnostics` to `master` with description "Add plan diff CLI command with ChangeSet rendering and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m4-plan-diagnostics`
- [ ] **COMMIT (Owner: Jeff | Group: A4c.artifacts | Branch: feature/m4-plan-diagnostics) - Commit message: "feat(cli): add plan artifacts and prompt commands"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m4-plan-diagnostics`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Implement `agents plan artifacts <plan_id>` to list stored artifacts (decision snapshots, diff, logs, validation summaries).
  - [ ] Code [Jeff]: Implement `agents plan prompt <plan_id> <guidance>` to append guidance and record a `prompt_definition` decision.
  - [ ] Code [Jeff]: Persist prompt guidance as plan metadata for follow-on execution and attach to decision tree.
  - [ ] Docs [Jeff]: Add CLI reference docs for `plan artifacts` and `plan prompt` with output field descriptions.
  - [ ] Tests (Behave) [Jeff]: Add scenarios for artifacts listing and prompt recording.
  - [ ] Tests (Robot) [Jeff]: Add Robot CLI tests covering artifacts and prompt commands.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/plan_artifacts_cli_bench.py` for listing overhead.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .`
  - [ ] Git [Jeff]: `git commit -m "feat(cli): add plan artifacts and prompt commands"`
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m4-plan-diagnostics` to `master` with description "Add plan artifacts + prompt CLI commands with persistence and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m4-plan-diagnostics`

**Parallel Group A7: Session Management [Jeff + Luis + Brent]** (M3; post-M1; depends on actor registry + plan lifecycle)
  **PARALLEL SUBTRACK A7.domain [Jeff]**: Session domain models + service contracts
  **PARALLEL SUBTRACK A7.persistence [Luis]**: DB tables + repositories + service implementation
  **PARALLEL SUBTRACK A7.cli [Brent]**: Session CLI commands + output formatting
  **SEQUENTIAL NOTE**: A7.domain must land before A7.persistence/cli; A7.cli depends on A7.persistence service wiring.
- [ ] **COMMIT (Owner: Jeff | Group: A7.domain | Branch: feature/m3-session-domain) - Commit message: "feat(session): add session domain models and contracts"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-session-domain`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Add `Session` and `SessionMessage` domain models with ULID IDs, actor ref, timestamps, and role enum (user/assistant/system/tool).
  - [ ] Code [Jeff]: Add message validation for role/content presence and stable ordering by sequence number.
  - [ ] Code [Jeff]: Add `SessionService` interface contracts (create/list/show/delete/append/export/import) with error types.
  - [ ] Docs [Jeff]: Add `docs/reference/session_model.md` and `docs/reference/session_service.md`.
  - [ ] Tests (Behave) [Jeff]: Add `features/session_model.feature` for model validation and serialization ordering.
  - [ ] Tests (Robot) [Jeff]: Add `robot/session_model.robot` smoke tests.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/session_model_bench.py` for validation throughput.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .`
  - [ ] Git [Jeff]: `git commit -m "feat(session): add session domain models and contracts"`
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-session-domain` to `master` with description "Add session domain models + service contracts with docs and tests."
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-session-domain`
- [ ] **COMMIT (Owner: Luis | Group: A7.persistence | Branch: feature/m3-session-persistence) - Commit message: "feat(session): add session persistence and repositories"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m3-session-persistence`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Add DB tables `sessions` and `session_messages` with indexes on session_id, created_at, and actor_name.
  - [ ] Code [Luis]: Implement SessionRepository + SessionMessageRepository with pagination and message append semantics.
  - [ ] Code [Luis]: Implement SessionService concrete class with export/import JSON payloads and truncation guards.
  - [ ] Code [Luis]: Wire session repositories/services into DI container and settings.
  - [ ] Docs [Luis]: Update `docs/reference/database_schema.md` with session tables and constraints.
  - [ ] Tests (Behave) [Luis]: Add `features/session_persistence.feature` for create/list/show/delete/append/export/import.
  - [ ] Tests (Robot) [Luis]: Add `robot/session_persistence.robot` integration smoke tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/session_persistence_bench.py` for write/read throughput.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .`
  - [ ] Git [Luis]: `git commit -m "feat(session): add session persistence and repositories"`
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m3-session-persistence` to `master` with description "Add session persistence tables + repositories + service wiring with tests."
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m3-session-persistence`
- [ ] **COMMIT (Owner: Brent | Group: A7.cli | Branch: feature/m3-session-cli) - Commit message: "feat(cli): add session commands"**
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git pull origin master`
  - [ ] Git [Brent]: `git checkout -b feature/m3-session-cli`
  - [ ] Git [Brent]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Brent]: Implement `agents session create/list/show/delete/export/import/tell` with `--actor`, `--stream`, and format outputs.
  - [ ] Code [Brent]: Ensure `session tell` uses SessionService append + actor execution, updates last_active, and persists message order.
  - [ ] Code [Brent]: Validate `session export/import` paths (create dirs, refuse overwrite unless `--force`).
  - [ ] Code [Brent]: Add `--format json/yaml` support for list/show and include message counts + last_active timestamps.
  - [ ] Docs [Brent]: Update CLI reference with session command examples, streaming notes, and JSON output shape.
  - [ ] Tests (Behave) [Brent]: Add `features/session_cli.feature` covering create/list/show/delete/export/import/tell flows + error cases.
  - [ ] Tests (Robot) [Brent]: Add `robot/session_cli.robot` end-to-end session flows with export/import round-trip.
  - [ ] Tests (ASV) [Brent]: Add `asv/benchmarks/session_cli_bench.py` for command parsing overhead.
  - [ ] Quality [Brent]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Brent]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Brent]: `git add .`
  - [ ] Git [Brent]: `git commit -m "feat(cli): add session commands"`
  - [ ] Forgejo PR [Brent]: Open PR from `feature/m3-session-cli` to `master` with description "Add session CLI commands with Behave/Robot coverage and docs updates."
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git branch -d feature/m3-session-cli`

**Parallel Group A8: Config CLI [Brent]** (M3; post-M1; depends on Settings)
- [ ] **COMMIT (Owner: Brent | Group: A8.cli | Branch: feature/m3-config-cli) - Commit message: "feat(cli): add config get/set/list commands"**
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git pull origin master`
  - [ ] Git [Brent]: `git checkout -b feature/m3-config-cli`
  - [ ] Git [Brent]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Brent]: Implement `agents config set <key> <value>`, `agents config get <key>`, and `agents config list [<regex>]` with `--filter-values` regex support.
  - [ ] Code [Brent]: Ensure config updates write to the correct config file path; create config file if missing.
  - [ ] Code [Brent]: Mask secret values in `config list` output by default with `--show-secrets` override.
  - [ ] Code [Brent]: Emit explicit errors for unknown keys and invalid regex filters.
  - [ ] Docs [Brent]: Add `docs/reference/config_cli.md` with examples and masking rules for secrets.
  - [ ] Tests (Behave) [Brent]: Add `features/config_cli.feature` for set/get/list, filter behavior, and secret masking.
  - [ ] Tests (Robot) [Brent]: Add `robot/config_cli.robot` smoke tests for config list output.
  - [ ] Tests (ASV) [Brent]: Add `asv/benchmarks/config_cli_bench.py` for CLI parsing.
  - [ ] Quality [Brent]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Brent]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Brent]: `git add .`
  - [ ] Git [Brent]: `git commit -m "feat(cli): add config get/set/list commands"`
  - [ ] Forgejo PR [Brent]: Open PR from `feature/m3-config-cli` to `master` with description "Add config get/set/list CLI commands with tests and docs updates."
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git branch -d feature/m3-config-cli`

**M1 SUCCESS CRITERIA (Day 7 MVP - source code only)**:
- Action created from YAML config and persisted (namespaced name, invariants, automation profile).
- Project created and linked to a local git-checkout resource.
- Plan use -> strategize -> execute -> apply completes with sandbox isolation and diff review.
- Tool-based change tracking produces a ChangeSet and applies to the repo after approval.
- `nox` passes with coverage >=97% on the MVP end-to-end path.

---

### Section 4: Projects & Resources [WORKSTREAM B - Hamza Lead]

**Target: Milestone M2 (+10 days)**
**Week 1-2 focus**: local source code only (git-checkout + fs-directory). Database, API, and remote resources are schema-only stubs for future work.

**Parallel Group B1: Resource Registry Core [Hamza + Jeff]** (can start after A5.alpha migrations are available)
  **SEQUENTIAL NOTE**: B1 domain models can start immediately; B1 DB migrations must rebase on the latest Alembic head after A5.alpha to keep a linear migration chain.
- [ ] **COMMIT (Owner: Hamza | Group: B1.core | Branch: feature/m2-resource-core-models) - Commit message: "feat(domain): add resource type spec and resource model"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m2-resource-core-models`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Create `src/cleveragents/domain/models/core/resource_type.py` with `ResourceTypeSpec`, `ResourceTypeArgument`, `ResourceKind` (physical/virtual), and `SandboxStrategy` enum.
  - [ ] Code [Hamza]: Define `ResourceTypeArgument` fields for CLI flags (`flag`, `help`, `arg_type`, `required`, `default`, `repeatable`, `choices`) and validate flag naming conventions.
  - [ ] Code [Hamza]: Add resource type fields: `user_addable`, `allowed_parents`, `allowed_children`, `auto_discover`, `handler` reference, and `sandbox_strategy` default.
  - [ ] Code [Hamza]: Add `Resource` and `ResourceRef` models with ULID, optional namespaced name, type name, location, description, sandbox strategy, read_only, and metadata.
  - [ ] Code [Hamza]: Enforce user-added resources require namespaced name; auto-discovered child resources are ULID-only with no namespaced name.
  - [ ] Code [Hamza]: Add virtual resource validation rules (no location; cannot be directly read/write; distinct from physical).
  - [ ] Code [Hamza]: Add validators for namespaced naming, ULID format, and parent/child DAG sanity (no self loops, no duplicate edges, type compatibility).
  - [ ] Code [Hamza]: Add `ResourceName.parse()` and `ResourceTypeName.parse()` helpers to normalize `local/` default and reject invalid namespaces.
  - [ ] Code [Hamza]: Add `docs/schema/resource_type.schema.yaml` with CLI argument definitions, parent/child constraints, and handler metadata.
  - [ ] Code [Hamza]: Add resource type YAML loader in `src/cleveragents/resource/schema.py` with version guard and clear error messages.
  - [ ] Docs [Hamza]: Add `docs/reference/resource_model.md` with examples for git-checkout and fs-directory resources plus physical/virtual notes.
  - [ ] Tests (Behave) [Hamza]: Add scenarios validating ULID format, namespace rules, allowed parent/child type checks, and sandbox strategy defaults.
  - [ ] Tests (Robot) [Hamza]: Add Robot test that loads a ResourceTypeSpec YAML fixture and validates it.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/resource_model_bench.py` for resource validation + DAG checks.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .`
  - [ ] Git [Hamza]: `git commit -m "feat(domain): add resource type spec and resource model"`
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m2-resource-core-models` to `master` with description "Add resource type spec + resource model with validation, schema, and tests for M2.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m2-resource-core-models`
- [ ] **COMMIT (Owner: Hamza | Group: B1.core | Branch: feature/m2-resource-core-models) - Commit message: "feat(resource): add built-in physical resource type configs"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m2-resource-core-models`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add built-in **physical** resource type YAML configs under `resources/types/` (M2 minimum set for source code only):
    - `git-checkout` (user-addable; `--path`, optional `--branch`; default sandbox_strategy = `git_worktree`)
    - `fs-mount` (user-addable; `--path`, optional `--read-only`)
    - `fs-directory` + `fs-file` (auto-discovered children of `fs-mount`/`git-checkout`; not user-addable)
  - [ ] Code [Hamza]: Include `schema_version`, `resource_kind: physical`, `sandbox_strategy`, and full allowed parent/child constraints per spec.
  - [ ] Code [Hamza]: Define CLI argument specs for user-addable types (`git`, `git-checkout`, `fs-mount`, `fs-directory`) including `--path`, `--url`, `--branch`, `--mount-path`, `--read-only` as applicable.
  - [ ] Code [Hamza]: Encode `auto_discover` rules for git and filesystem trees (git children + fs-directory children) with bounded depth and type filters.
  - [ ] Code [Hamza]: Add bootstrap registration in `ResourceRegistryService` (register physical built-ins on startup if missing; idempotent).
  - [ ] Docs [Hamza]: Update `docs/reference/resource_types_builtin.md` with physical type flags, parents/children, and examples.
  - [ ] Tests (Behave) [Hamza]: Add scenarios ensuring physical built-ins exist and parent/child constraints are enforced.
  - [ ] Tests (Robot) [Hamza]: Add Robot test that lists resource types and asserts physical built-ins are present.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/resource_type_bootstrap_bench.py` for registration overhead.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .`
  - [ ] Git [Hamza]: `git commit -m "feat(resource): add built-in physical resource type configs"`
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m2-resource-core-models` to `master` with description "Add built-in physical resource type configs (git + fs) with discovery rules and docs.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m2-resource-core-models`
- [ ] **COMMIT (Owner: Hamza | Group: B1.core | Branch: feature/m2-resource-core-models) - Commit message: "feat(resource): add built-in virtual resource type configs"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m2-resource-core-models`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add built-in **virtual** resource type YAML configs under `resources/types/` (M2 minimum set): `file`, `directory`, `commit`, `branch`, `tag`, `tree`.
  - [ ] Code [Hamza]: Set `resource_kind: virtual`, `user_addable: false`, and no sandbox strategy; encode allowed children per spec.
  - [ ] Code [Hamza]: Add metadata for equivalence criteria (content hash/name/target) to support future auto-linking.
  - [ ] Code [Hamza]: Extend bootstrap registration to include virtual types and hide them from `resource add` subcommand generation.
  - [ ] Docs [Hamza]: Update `docs/reference/resource_types_builtin.md` with virtual type descriptions and link semantics.
  - [ ] Tests (Behave) [Hamza]: Add scenarios ensuring virtual built-ins exist and cannot be user-added.
  - [ ] Tests (Robot) [Hamza]: Add Robot test that lists resource types and asserts virtual built-ins are present.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/resource_type_virtual_bench.py` for registry performance.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .`
  - [ ] Git [Hamza]: `git commit -m "feat(resource): add built-in virtual resource type configs"`
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m2-resource-core-models` to `master` with description "Add built-in virtual resource type configs and registration rules.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m2-resource-core-models`
- [ ] **COMMIT (Owner: Hamza | Group: B1.core | Branch: feature/m2-resource-core-models) - Commit message: "feat(domain): add project model v3 with linked resources"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m2-resource-core-models`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add `Project`, `ProjectResourceLink`, `ProjectValidationSummary` (derived from validation attachments), and `ProjectContextPolicy` models using namespaced name as the unique identifier (no ULID per spec).
  - [ ] Code [Hamza]: Add fields for `invariants`, `invariant_actor`, `automation_profile`, and `context_views` (strategize/execute/apply/default).
  - [ ] Code [Hamza]: Add validation for resource link overrides (read_only flags, alias uniqueness, resource existence).
  - [ ] Code [Hamza]: Add helpers to compute effective invariants and automation profile (project defaults).
  - [ ] Code [Hamza]: Define default context view policy when `context_views` is omitted (inherit from `default` view; explicit override per view).
  - [ ] Code [Hamza]: Enforce non-empty invariant text and deterministic ordering for project-level invariants.
  - [ ] Docs [Hamza]: Add `docs/reference/project_model.md` describing resource linking, validation attachments, and context view policies.
  - [ ] Tests (Behave) [Hamza]: Add scenarios for project model validation, link overrides, and context view inheritance.
  - [ ] Tests (Robot) [Hamza]: Add Robot test that creates a Project object and prints serialized output.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/project_model_bench.py` for serialization/validation performance.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .`
  - [ ] Git [Hamza]: `git commit -m "feat(domain): add project model v3 with linked resources"`
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m2-resource-core-models` to `master` with description "Add project model v3 with resource links, invariants, and context view policy fields.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m2-resource-core-models`
- [ ] **COMMIT (Owner: Jeff | Group: B1.core | Branch: feature/m2-resource-core-db) - Commit message: "feat(db): add resource registry tables"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m2-resource-core-db`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Add Alembic migration for `resource_types`, `resources`, and `resource_edges` tables with naming conventions.
  - [ ] Code [Jeff]: Define `resource_types` columns: `name`, `namespace`, `description`, `resource_kind`, `sandbox_strategy`, `user_addable`, `handler_ref`, `args_schema_json`, `allowed_parent_types_json`, `allowed_child_types_json`, `auto_discover_json`, timestamps.
  - [ ] Code [Jeff]: Define `resources` columns: ULID PK, `namespaced_name`, `namespace`, `type_name`, `resource_kind`, `location`, `description`, `read_only`, `metadata_json`, `sandbox_strategy`, timestamps.
  - [ ] Code [Jeff]: Add check constraints for `resource_kind` enum values and enforce `namespaced_name` uniqueness when non-null.
  - [ ] Code [Jeff]: Add FK from `resources.type_name` -> `resource_types.name` with restrict-on-delete to prevent orphaned types.
  - [ ] Code [Jeff]: Define `resource_edges` columns: `parent_id`, `child_id`, `created_at`, with uniqueness constraint and FK cascade rules.
  - [ ] Code [Jeff]: Add indexes on `resources.namespaced_name`, `resources.namespace`, `resources.type_name`, and `resource_edges.parent_id/child_id`.
  - [ ] Docs [Jeff]: Update `docs/reference/database_schema.md` with resource registry tables and constraints.
  - [ ] Tests (Behave) [Jeff]: Add migration scenarios verifying tables, indices, and edge uniqueness.
  - [ ] Tests (Robot) [Jeff]: Add Robot migration smoke test using `nox -s db_migrate`.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/resource_registry_migration_bench.py` for migration baseline.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .`
  - [ ] Git [Jeff]: `git commit -m "feat(db): add resource registry tables"`
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m2-resource-core-db` to `master` with description "Add resource registry tables, indexes, and migration tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m2-resource-core-db`

**Parallel Group B2: Project Persistence + Services [Hamza + Luis]** (depends on B1 domain models)
- [ ] **COMMIT (Owner: Jeff | Group: B2.persistence | Branch: feature/m2-projects-db) - Commit message: "feat(db): add projects and project links tables"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m2-projects-db`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Add Alembic migration skeleton with explicit down_revision to latest A5.alpha head.
  - [ ] Code [Jeff]: Define `projects` table with namespaced_name PK, namespace, description, automation_profile, invariant_actor, invariants_json, context_policy_json, tags_json, created_by, timestamps.
  - [ ] Code [Jeff]: Add `projects` constraints for non-empty names, namespace/name derivation consistency, and unique namespaced_name.
  - [ ] Code [Jeff]: Add explicit `context_policy_json` default (`{}`) and enforce JSON schema compatibility for future migrations.
  - [ ] Code [Jeff]: Define `project_resource_links` table with link_id ULID, project_name FK, resource_id FK, alias, read_only, created_at.
  - [ ] Code [Jeff]: Add uniqueness constraint on (project_name, resource_id) and index on (project_name, alias) for fast lookups.
  - [ ] Code [Jeff]: Add indexes on `project_resource_links.project_name` and `project_resource_links.resource_id` for joins.
  - [ ] Docs [Jeff]: Document project table schema and link semantics in `docs/reference/database_schema.md`.
  - [ ] Tests (Behave) [Jeff]: Add migration scenarios verifying project tables, FK constraints, and unique link enforcement.
  - [ ] Tests (Robot) [Jeff]: Add Robot test that inserts a project and link row and validates alias uniqueness.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/project_migration_bench.py` for migration baseline.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Commit [Jeff]: `git commit -m "feat(db): add projects and project links tables"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m2-projects-db` to `master` with description "Add projects + project_resource_links tables with constraints and migration tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m2-projects-db`
- [ ] **COMMIT (Owner: Hamza | Group: B2.persistence | Branch: feature/m2-resource-project-repos) - Commit message: "feat(repo): add resource repositories"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m2-resource-project-repos`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Implement `ResourceTypeRepository` CRUD and `ResourceRepository` CRUD with DAG edge helpers.
  - [ ] Code [Hamza]: Add methods for tree traversal, child discovery queries, and name/ULID resolution.
  - [ ] Code [Hamza]: Add repository guardrails for preventing cycles and duplicate edges.
  - [ ] Code [Hamza]: Add repository methods for `resolve_namespaced_name` and `resolve_ulid` with clear NotFound errors.
  - [ ] Docs [Hamza]: Document repository interfaces in `docs/reference/repositories.md`.
  - [ ] Tests (Behave) [Hamza]: Add repository scenarios for create/get/list/tree and cycle rejection.
  - [ ] Tests (Robot) [Hamza]: Add Robot test exercising tree output ordering.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/resource_repository_bench.py` for tree query performance.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Commit [Hamza]: `git commit -m "feat(repo): add resource repositories"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m2-resource-project-repos` to `master` with description "Add resource repositories with DAG helpers, docs, and tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m2-resource-project-repos`
- [ ] **COMMIT (Owner: Hamza | Group: B2.persistence | Branch: feature/m2-resource-project-repos) - Commit message: "feat(repo): add project repositories"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m2-resource-project-repos`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Implement `ProjectRepository` and `ProjectResourceLinkRepository` with namespace filtering and name-based lookup.
  - [ ] Code [Hamza]: Add methods to list project context policies and derived validation attachment summaries for linked resources.
  - [ ] Code [Hamza]: Enforce resource existence + type validation when linking (reject missing resource IDs early).
  - [ ] Docs [Hamza]: Update repository docs with project link examples and validation attachment notes.
  - [ ] Tests (Behave) [Hamza]: Add scenarios for project create/link/unlink and validation attachment summaries.
  - [ ] Tests (Robot) [Hamza]: Add Robot test that links two resources to one project.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/project_repository_bench.py` for link/unlink performance.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Commit [Hamza]: `git commit -m "feat(repo): add project repositories"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m2-resource-project-repos` to `master` with description "Add project repositories + link repositories with docs and tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m2-resource-project-repos`
- [ ] **COMMIT (Owner: Hamza | Group: B2.service | Branch: feature/m2-resource-registry-service) - Commit message: "feat(service): add resource registry service"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m2-resource-registry-service`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Implement `ResourceRegistryService` for register/remove/show/tree operations with name/ULID resolution.
  - [ ] Code [Hamza]: Add auto-discovery hook that delegates to resource handlers (git-checkout for MVP).
  - [ ] Code [Hamza]: Add validation that resource type supports parent/child linkage before linking.
  - [ ] Code [Hamza]: Ensure `register` returns the persisted resource with ULID + resolved namespaced name for CLI output consistency.
  - [ ] Docs [Hamza]: Add `docs/reference/resource_registry.md` describing API behavior and error cases.
  - [ ] Tests (Behave) [Hamza]: Add scenarios for register/remove/show/tree behavior and auto-discovery.
  - [ ] Tests (Robot) [Hamza]: Add Robot test that registers a git-checkout and inspects child count.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/resource_registry_service_bench.py` for register/show performance.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Commit [Hamza]: `git commit -m "feat(service): add resource registry service"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m2-resource-registry-service` to `master` with description "Add resource registry service with auto-discovery and tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m2-resource-registry-service`
- [ ] **COMMIT (Owner: Luis | Group: B2.service | Branch: feature/m2-project-service) - Commit message: "feat(service): add project service v3"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m2-project-service`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Implement `ProjectService` create/list/show/delete/link/unlink methods using repositories.
  - [ ] Code [Luis]: Add validation attachment helpers (read-only listing of validation attachments for linked resources) and context policy setters for project views.
  - [ ] Code [Luis]: Enforce read-only resource links and project-level invariant actor defaults.
  - [ ] Code [Luis]: Enforce namespaced project naming on create and reject duplicate names with explicit error details.
  - [ ] Docs [Luis]: Update `docs/reference/project_service.md` with usage examples and error cases.
  - [ ] Tests (Behave) [Luis]: Add scenarios for project create/link/unlink/context policy + validation attachment visibility.
  - [ ] Tests (Robot) [Luis]: Add Robot test that creates project and links a resource.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/project_service_bench.py` for link/unlink performance.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Commit [Luis]: `git commit -m "feat(service): add project service v3"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m2-project-service` to `master` with description "Add project service v3 with linking, invariants, and context policy helpers.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m2-project-service`

**Parallel Group B3: CLI Commands [Hamza]** (depends on B2 services)
- [ ] **COMMIT (Owner: Hamza | Group: B3.cli | Branch: feature/m2-project-resource-cli) - Commit message: "feat(cli): add resource type commands"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m2-project-resource-cli`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add `agents resource type add/remove/list/show` commands with YAML config input, schema validation, and clear field-path errors.
  - [ ] Code [Hamza]: Implement `--update` behavior with conflict detection and explicit `name` mismatch errors.
  - [ ] Code [Hamza]: Wire `resource type add` to `ResourceTypeSpec` loader with schema version guard and path resolution.
  - [ ] Code [Hamza]: Ensure `resource type list` supports `--type` and `--regex` filters with stable ordering + `--format json/yaml`.
  - [ ] Code [Hamza]: Include `resource_kind`, `sandbox_strategy`, `user_addable`, and `handler` in `resource type show` output.
  - [ ] Docs [Hamza]: Update CLI reference with resource type examples and expected output columns.
  - [ ] Tests (Behave) [Hamza]: Add scenarios for resource type lifecycle, update conflicts, and invalid schema handling.
  - [ ] Tests (Robot) [Hamza]: Add Robot suite `robot/resource_type_cli.robot`.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/resource_type_cli_bench.py` for config parsing overhead.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Commit [Hamza]: `git commit -m "feat(cli): add resource type commands"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m2-project-resource-cli` to `master` with description "Add resource type CLI commands with schema validation and tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m2-project-resource-cli`
- [ ] **COMMIT (Owner: Hamza | Group: B3.cli | Branch: feature/m2-project-resource-cli) - Commit message: "feat(cli): add resource commands"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m2-project-resource-cli`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add `agents resource add/remove/list/show/tree` commands with type-specific flags and name/ULID resolution.
  - [ ] Code [Hamza]: Implement dynamic option parsing for type-specific flags from ResourceTypeSpec (including repeatable args).
  - [ ] Code [Hamza]: Implement `resource inspect --tree/--file` per spec for resource introspection and path-scoped views.
  - [ ] Code [Hamza]: Add `resource link-child` and `resource unlink-child` commands with type compatibility validation.
  - [ ] Code [Hamza]: Ensure `resource show` includes `resource_kind`, `sandbox_strategy`, `read_only`, and parent/child counts.
  - [ ] Docs [Hamza]: Update CLI reference with resource examples (git-checkout, fs-directory) and output columns.
  - [ ] Tests (Behave) [Hamza]: Add scenarios for resource registration, list filters, tree rendering, and link-child constraints.
  - [ ] Tests (Robot) [Hamza]: Add Robot suite `robot/resource_cli.robot`.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/resource_cli_bench.py` for command parsing and list output.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Commit [Hamza]: `git commit -m "feat(cli): add resource commands"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m2-project-resource-cli` to `master` with description "Add resource CLI commands (add/show/list/tree/inspect/link-child) with tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m2-project-resource-cli`
- [ ] **COMMIT (Owner: Hamza | Group: B3.cli | Branch: feature/m2-project-resource-cli) - Commit message: "feat(cli): add project commands"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m2-project-resource-cli`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add `agents project create/show/list/delete/link-resource/unlink-resource` commands using namespaced project names.
  - [ ] Code [Hamza]: Implement `project create` flags for `--description`, `--resource`, `--invariant`, `--invariant-actor`, and `--automation-profile`.
  - [ ] Code [Hamza]: Ensure `project show` includes linked resources, read_only flags, aliases, and validation attachment summaries.
  - [ ] Code [Hamza]: Ensure `project list` includes resource counts, invariant counts, and automation profile summaries.
  - [ ] Docs [Hamza]: Update CLI reference with project examples and validation attachment visibility (via `agents validation attach`).
  - [ ] Tests (Behave) [Hamza]: Add scenarios for project create/link/unlink/show/list/delete and validation display.
  - [ ] Tests (Robot) [Hamza]: Add Robot suite `robot/project_cli.robot`.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/project_cli_bench.py` for command parsing and list output.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Commit [Hamza]: `git commit -m "feat(cli): add project commands"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m2-project-resource-cli` to `master` with description "Add project CLI commands (create/show/list/link/unlink) with validation attachment summaries.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m2-project-resource-cli`

**Parallel Group B3.context: Project Context Views [Hamza]** (depends on B1 project model + B2 services; post-M2)
  **PARALLEL SUBTRACK B3.context.service [Hamza]**: Context policy service + persistence helpers
  **PARALLEL SUBTRACK B3.context.cli [Hamza]**: CLI commands for context set/show/inspect/simulate
  **SEQUENTIAL NOTE**: B3.context.service must land before B3.context.cli to lock policy serialization and defaults.
- [ ] **COMMIT (Owner: Hamza | Group: B3.context.service | Branch: feature/m3-project-context-policy) - Commit message: "feat(service): add project context policy service"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m3-project-context-policy`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Implement ProjectContextPolicy serialization (include/exclude resources, include/exclude paths, token budgets, summarize flags, summary max tokens, strategy list, depth gradients, skeleton ratio).
  - [ ] Code [Hamza]: Add ProjectContextService helpers to set/clear/show/inspect policy and compute effective view defaults.
  - [ ] Code [Hamza]: Add validation for invalid resource refs, conflicting include/exclude, and max token bounds.
  - [ ] Docs [Hamza]: Add `docs/reference/project_context_policy.md` with flag mapping and defaults.
  - [ ] Tests (Behave) [Hamza]: Add scenarios for policy validation, defaults, and serialization round-trip.
  - [ ] Tests (Robot) [Hamza]: Add Robot smoke tests for policy persistence.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/project_context_policy_bench.py` for policy parse overhead.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Commit [Hamza]: `git commit -m "feat(service): add project context policy service"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m3-project-context-policy` to `master` with description "Add project context policy service with serialization, defaults, and tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m3-project-context-policy`
- [ ] **COMMIT (Owner: Hamza | Group: B3.context.cli | Branch: feature/m3-project-context-cli) - Commit message: "feat(cli): add project context set/show/inspect/simulate"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m3-project-context-cli`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Implement `agents project context set/show/inspect/simulate` with full flag set per spec (view, include/exclude resource, include/exclude path, budget limits, default breadth/depth, depth gradients, temporal scope, auto-refresh, summarize flags, summary max tokens, strategy list).
  - [ ] Code [Hamza]: Add validation for mutually exclusive flags and display `--clear` behavior explicitly in output.
  - [ ] Code [Hamza]: Ensure `context inspect` shows resolved policy + derived budgets; `context simulate` prints estimated payload size + token budget allocation.
  - [ ] Docs [Hamza]: Update CLI reference with context command examples and default behaviors.
  - [ ] Tests (Behave) [Hamza]: Add scenarios for set/show/inspect/simulate including invalid flags and default view inheritance.
  - [ ] Tests (Robot) [Hamza]: Add `robot/project_context_cli.robot` for context commands.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/project_context_cli_bench.py` for command parsing.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Commit [Hamza]: `git commit -m "feat(cli): add project context set/show/inspect/simulate"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m3-project-context-cli` to `master` with description "Add project context CLI commands (set/show/inspect/simulate) with tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m3-project-context-cli`

**Parallel Group B3.cleanup: Legacy Project Removal [Jeff]** (after B3.cli lands)

- [ ] **COMMIT (Owner: Jeff | Group: B3.cleanup | Branch: feature/m2-project-legacy-cleanup) - Commit message: "refactor(project): remove legacy project init/status commands"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m2-project-legacy-cleanup`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Remove legacy `agents project init/status/clean/file-filter` commands from `src/cleveragents/cli/commands/project.py`.
  - [ ] Code [Jeff]: Remove or alias legacy `agents context` commands to `agents project context` to match spec surface.
  - [ ] Code [Jeff]: Remove `.cleveragents` directory bootstrap logic from legacy `ProjectService` and deprecate `ProjectSettings` fields tied to local init.
  - [ ] Code [Jeff]: Update `src/cleveragents/application/container.py` to stop wiring legacy ProjectService once v3 service is in place.
  - [ ] Code [Jeff]: Remove legacy `src/cleveragents/domain/models/core/project.py` in favor of v3 project model and update imports.
  - [ ] Docs [Jeff]: Remove references to `agents project init` from CLI docs and point to `agents project create` + `agents init` (global) flows.
  - [ ] Tests (Behave) [Jeff]: Remove/replace legacy project init scenarios with v3 project create scenarios.
  - [ ] Tests (Robot) [Jeff]: Remove legacy project init Robot suites and add v3 replacements if missing.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/project_cli_cleanup_bench.py` for CLI help/rendering baseline after removal.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Commit [Jeff]: `git commit -m "refactor(project): remove legacy project init/status commands"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m2-project-legacy-cleanup` to `master` with description "Remove legacy project init/status commands and align CLI to v3 project flows.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m2-project-legacy-cleanup`

**Parallel Group B4: Sandboxing [Luis + Jeff]** (depends on resource registry + project links)
- [ ] **COMMIT (Owner: Luis | Group: B4.sandbox | Branch: feature/m2-sandbox-core) - Commit message: "feat(sandbox): add sandbox strategy interface and manager"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m2-sandbox-core`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [X] Code [Luis]: Add `SandboxStrategy` protocol, `SandboxRef`, `SandboxManager`, and `SandboxRegistry` with per-resource sandboxes. (protocol, factory, manager, NoSandbox, merge strategies, status enum all implemented)
  - [X] Code [Luis]: Implement lazy sandbox creation, cleanup hooks, and plan-scoped retention policy stubs. (completed by Luis)
  - [ ] Code [Luis]: Add sandbox path rewriting helper for tool execution and MCP adapters.
  - [ ] Code [Luis]: Include `resource_id`, `plan_id`, and `sandbox_path` in `SandboxRef` for traceability in logs.
  - [ ] Docs [Luis]: Add `docs/reference/sandbox.md` describing lifecycle, APIs, and path rewriting rules.
  - [ ] Tests (Behave) [Luis]: Add scenarios for sandbox manager creation, cleanup, and path rewrite behavior.
  - [ ] Tests (Robot) [Luis]: Add Robot test that creates a sandbox and verifies filesystem isolation.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/sandbox_manager_bench.py` for sandbox creation overhead.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Commit [Luis]: `git commit -m "feat(sandbox): add sandbox strategy interface and manager"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m2-sandbox-core` to `master` with description "Add sandbox strategy interface, manager, and docs/tests for path rewriting.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m2-sandbox-core`
- [ ] **COMMIT (Owner: Luis | Group: B4.sandbox | Branch: feature/m2-sandbox-core) - Commit message: "feat(sandbox): implement git_worktree strategy"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m2-sandbox-core`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Implement git worktree creation, checkout, and cleanup for git-checkout resources.
  - [ ] Code [Luis]: Add safe fallback for repositories without clean worktrees and clear error messages.
  - [ ] Code [Luis]: Record sandbox metadata (worktree path, branch, base commit) for rollback.
  - [ ] Code [Luis]: Name worktree branches using plan ULID to avoid collisions across concurrent plans.
  - [ ] Docs [Luis]: Update sandbox doc with git_worktree usage and rollback behavior.
  - [ ] Tests (Behave) [Luis]: Add scenarios for git worktree sandbox creation and rollback.
  - [ ] Tests (Robot) [Luis]: Add Robot test that modifies sandbox and verifies original repo unchanged.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/git_worktree_bench.py` for sandbox creation time.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Commit [Luis]: `git commit -m "feat(sandbox): implement git_worktree strategy"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m2-sandbox-core` to `master` with description "Implement git_worktree sandbox strategy with rollback metadata and tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m2-sandbox-core`
- [ ] **COMMIT (Owner: Hamza | Group: B4.sandbox | Branch: feature/m2-resource-git-handler) - Commit message: "feat(resource): add git-checkout handler and discovery"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m2-resource-git-handler`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add git-checkout handler that validates repo path, branch, and read_only flags.
  - [ ] Code [Hamza]: Implement child resource discovery for fs-directory children (schema-only for now) and record ULID-only children.
  - [ ] Code [Hamza]: Add sandbox strategy mapping for git-checkout and path normalization helpers.
  - [ ] Code [Hamza]: Return explicit error when repo path is missing or not a git repo (include path in message).
  - [ ] Docs [Hamza]: Document git-checkout handler behavior in `docs/reference/resources_git.md`.
  - [ ] Tests (Behave) [Hamza]: Add scenarios for handler validation and discovery counts.
  - [ ] Tests (Robot) [Hamza]: Add Robot test registering a git repo and asserting discovered children.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/git_discovery_bench.py` for discovery cost.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Commit [Hamza]: `git commit -m "feat(resource): add git-checkout handler and discovery"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m2-resource-git-handler` to `master` with description "Add git-checkout resource handler with discovery and tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m2-resource-git-handler`
- [ ] **COMMIT (Owner: Luis | Group: B4.sandbox | Branch: feature/m2-sandbox-core) - Commit message: "feat(sandbox): add copy_on_write strategy stub"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m2-sandbox-core`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Add copy_on_write strategy skeleton with TODOs for large-project optimization.
  - [ ] Code [Luis]: Raise explicit NotImplementedError with guidance on when it will be available.
  - [ ] Docs [Luis]: Document that copy_on_write is stubbed for post-M1 work.
  - [ ] Docs [Luis]: Add a CLI note in `docs/reference/sandbox.md` showing the exact error message returned by the stub.
  - [ ] Tests (Behave) [Luis]: Add scenario that selecting copy_on_write raises NotImplementedError with clear message.
  - [ ] Tests (Robot) [Luis]: Add Robot test verifying stub error output.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/sandbox_stub_bench.py` (baseline no-op).
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Commit [Luis]: `git commit -m "feat(sandbox): add copy_on_write strategy stub"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m2-sandbox-core` to `master` with description "Add copy_on_write sandbox strategy stub with explicit NotImplementedError and tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m2-sandbox-core`

**M2 MERGE GATE**:
- Register a git-checkout resource and link it to a project via CLI.
- Create a sandbox for the linked resource and verify isolation via tests.
- Project context commands and validation attachment visibility work and persist.
- `nox` passes with coverage >=97%.

**M2 SUCCESS CRITERIA**:
- Resource registry supports resource types, resources, and DAG links with persistence (tables + repositories).
- Projects can link/unlink resources with CLI commands for resource types/resources/projects (list/show/tree included).
- Validation attachments (via `agents validation attach/detach`) appear in `project show` outputs.
- Git-checkout sandbox isolates changes; copy_on_write strategy returns clear NotImplementedError for fs-directory (documented).
- Resource/project services are DI-wired and exercised by Behave + Robot suites.
- `nox` passes with coverage >=97% across resource/project suites.
---

### Section 5: Actors, Skills & Tool Execution [WORKSTREAM C - Aditya Lead]

**Target: Milestone M3 (+14 days)**

**Week 2 focus**: Actor YAML, compilation, skills, and tool-based change tracking.

**Parallel Group C0: Tool Registry + Validation System [Jeff + Luis]** (start Day 5; precedes C1/C3)
  **PARALLEL SUBTRACK C0.domain [Jeff]**: Tool + Validation domain models + schemas
  **PARALLEL SUBTRACK C0.registry [Luis]**: Tool registry persistence + repositories
  **PARALLEL SUBTRACK C0.cli [Jeff]**: CLI commands for tools/validations
  **SEQUENTIAL MERGE NOTE**: C0.domain must land before C0.registry/cli; C0.registry before C3 context wiring. C0.registry migrations must rebase after A5.alpha; C0.binding should wait for B1.core resource type constraints to validate bindings.
- [ ] **COMMIT (Owner: Jeff | Group: C0.domain | Branch: feature/m3-tool-domain) - Commit message: "feat(tool): add tool and validation domain models"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-tool-domain`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Add `Tool` model in `src/cleveragents/domain/models/core/tool.py` with namespaced name, description, source type, input/output JSON schema, and capability metadata (`read_only`, `writes`, `checkpointable`, `side_effects`).
  - [ ] Code [Jeff]: Enforce capability consistency (`read_only` implies `writes=false`; validation of conflicting flags).
  - [ ] Code [Jeff]: Add `ResourceBinding` model with slot definitions, binding modes (context, static, parameter), and required `access` level (`read_only`/`read_write`).
  - [ ] Code [Jeff]: Add `Validation` model as Tool subtype with `mode`, `wraps`, and `transform` fields; enforce read-only constraints.
  - [ ] Code [Jeff]: Enforce Validation output schema includes `passed: bool` plus optional `data`/`message`, and force `writes=false`, `checkpointable=false`, `read_only=true` for validations.
  - [ ] Code [Jeff]: Enforce namespace collision rules (a Validation and Tool cannot share the same namespaced name).
  - [ ] Code [Jeff]: Require `transform` when `wraps` is set and validate that `wraps` references an existing Tool name.
  - [ ] Code [Jeff]: Add enums for ToolSource, ToolType (tool/validation), and ValidationMode.
  - [ ] Code [Jeff]: Add `docs/schema/tool.schema.yaml` and `docs/schema/validation.schema.yaml` with required fields, `wraps`/`transform` rules, and resource binding definitions.
  - [ ] Code [Jeff]: Add YAML loader in `src/cleveragents/tool/schema.py` that validates schema version, normalizes keys, and returns Tool/Validation domain models.
  - [ ] Code [Jeff]: Add example configs under `examples/tools/` and `examples/validations/` (plain tool, validation, wrapped validation) for tests.
  - [ ] Docs [Jeff]: Add `docs/reference/tool_model.md` and `docs/reference/validation_model.md` with examples.
  - [ ] Tests (Behave) [Jeff]: Add `features/tool_model.feature` for schema validation, resource binding rules, validation constraints, and YAML loader errors.
  - [ ] Tests (Robot) [Jeff]: Add `robot/tool_model.robot` smoke tests for model creation.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/tool_model_bench.py` for schema validation throughput (model + YAML loader).
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(tool): add tool and validation domain models"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-tool-domain` to `master` with description "Add tool + validation domain models, schema loaders, and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-tool-domain`
- [ ] **COMMIT (Owner: Luis | Group: C0.registry | Branch: feature/m3-tool-registry) - Commit message: "feat(tool): add tool registry persistence"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m3-tool-registry`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Add DB tables for `tools`, `tool_bindings`, and `validation_attachments` with indexes on namespaced name and type.
  - [ ] Code [Luis]: Define `tools` columns: `namespaced_name`, `namespace`, `tool_type`, `source`, `description`, `input_schema_json`, `output_schema_json`, `capability_json`, `metadata_json`, `yaml_text`, timestamps.
  - [ ] Code [Luis]: Define `tool_bindings` columns: `binding_id` ULID, `tool_name`, `slot_name`, `binding_mode`, `resource_type`, `required`, `static_resource_id`, `static_resource_name`, timestamps.
  - [ ] Code [Luis]: Define `validation_attachments` columns: `attachment_id` ULID, `validation_name`, `resource_id`, optional `project_name`, optional `plan_id`, `args_json`, timestamps.
  - [ ] Code [Luis]: Enforce cross-type name uniqueness (tool vs validation) at persistence and service layers.
  - [ ] Code [Luis]: Enforce validation attachment scope rules (always resource-bound; optional project/plan only).
  - [ ] Code [Luis]: Persist original YAML text for auditability and re-render it on `tool show` output.
  - [ ] Code [Luis]: Add uniqueness constraints for `tools.namespaced_name` and `tool_bindings(tool_name, slot_name)`; index `validation_attachments.resource_id`.
  - [ ] Code [Luis]: Implement ToolRepository + ValidationAttachmentRepository with list/show filters and eager-loading of bindings.
  - [ ] Code [Luis]: Add ToolRegistryService for register/update/remove/list/show with name conflict checks and tool/validation type enforcement.
  - [ ] Docs [Luis]: Update `docs/reference/database_schema.md` with tool/validation tables.
  - [ ] Tests (Behave) [Luis]: Add `features/tool_registry.feature` for register/update/remove and validation-only constraints.
  - [ ] Tests (Robot) [Luis]: Add `robot/tool_registry.robot` for list/show smoke tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/tool_registry_bench.py` for registry list performance.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(tool): add tool registry persistence"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m3-tool-registry` to `master` with description "Add tool registry persistence with bindings and validation attachments.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m3-tool-registry`
- [ ] **COMMIT (Owner: Jeff | Group: C0.binding | Branch: feature/m3-tool-binding) - Commit message: "feat(tool): add resource binding resolution"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-tool-binding`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Implement binding resolution for contextual, static, and parameter bindings with type compatibility checks.
  - [ ] Code [Jeff]: Add resolution helpers for resource name/ULID lookup and project-scoped filtering.
  - [ ] Docs [Jeff]: Add `docs/reference/tool_bindings.md` with resolution order and examples.
  - [ ] Tests (Behave) [Jeff]: Add binding resolution scenarios (context vs static vs parameter).
  - [ ] Tests (Robot) [Jeff]: Add Robot test resolving a bound resource by name.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/binding_resolution_bench.py` for resolution latency.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(tool): add resource binding resolution"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-tool-binding` to `master` with description "Add tool resource binding resolution and docs/tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-tool-binding`
- [ ] **COMMIT (Owner: Jeff | Group: C0.cli | Branch: feature/m3-tool-cli) - Commit message: "feat(cli): add tool and validation commands"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-tool-cli`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Implement `agents tool add/remove/list/show` with YAML config input, schema validation, and `--type` filter.
  - [ ] Code [Jeff]: Add `--update` to `tool add` and emit conflict errors for name/type mismatches.
  - [ ] Code [Jeff]: Implement `agents validation add/attach/detach` commands and enforce validation-only name use.
  - [ ] Code [Jeff]: Support `validation attach --project/--plan` flags and store attachment args (JSON-serialized).
  - [ ] Code [Jeff]: Add `--format json/yaml` output for tool/validation list/show to align with schema fields.
  - [ ] Docs [Jeff]: Update CLI reference with tool/validation commands, attach examples, and output format details.
  - [ ] Tests (Behave) [Jeff]: Add CLI scenarios for tool/validation registration, update conflict, and attachment lifecycle.
  - [ ] Tests (Robot) [Jeff]: Add Robot CLI suites for tool and validation commands (add/show/attach/detach).
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/tool_cli_bench.py` for CLI parsing overhead.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(cli): add tool and validation commands"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-tool-cli` to `master` with description "Add tool/validation CLI commands with attachment workflows and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-tool-cli`

**Parallel Group C0.skill: Skill Registry & YAML [Aditya + Jeff + Luis]** (depends on C0.domain + C0.registry; must land before C3.protocol)
  **PARALLEL SUBTRACK C0.skill.schema [Aditya]**: Skill YAML schema + examples
  **PARALLEL SUBTRACK C0.skill.domain [Jeff]**: Skill domain model + resolver
  **PARALLEL SUBTRACK C0.skill.registry [Luis]**: Skill persistence + service
  **PARALLEL SUBTRACK C0.skill.cli [Aditya]**: CLI commands + output formatting
  **SEQUENTIAL MERGE NOTE**: C0.skill.domain must land before C0.skill.registry/C0.skill.cli to avoid dual representations.
- [ ] **COMMIT (Owner: Aditya | Group: C0.skill.schema | Branch: feature/m3-skill-schema) - Commit message: "docs(skill): add skill yaml schema and examples"**
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git pull origin master`
  - [ ] Git [Aditya]: `git checkout -b feature/m3-skill-schema`
  - [ ] Git [Aditya]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Docs [Aditya]: Author `docs/schema/skill.schema.yaml` with versioning, required fields, and explicit type constraints for tool refs, inline tools, includes, and MCP sources.
  - [ ] Docs [Aditya]: Add skill YAML examples under `examples/skills/` (single-tool, composed, inline tool, validation-only, MCP-backed).
  - [ ] Code [Aditya]: Add schema loader in `src/cleveragents/skills/schema.py` that validates schema version, normalizes keys, and returns typed data.
  - [ ] Code [Aditya]: Add clear validation errors for missing tools, recursive includes, and invalid namespaced names.
  - [ ] Tests (Behave) [Aditya]: Add `features/skill_schema.feature` scenarios validating each example and invalid cases.
  - [ ] Tests (Robot) [Aditya]: Add `robot/skill_schema.robot` to load and validate every example.
  - [ ] Tests (ASV) [Aditya]: Add `asv/benchmarks/skill_schema_bench.py` for schema validation throughput.
  - [ ] Quality [Aditya]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Aditya]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Aditya]: `git add .` (only after coverage check passes)
  - [ ] Commit [Aditya]: `git commit -m "docs(skill): add skill yaml schema and examples"`.
  - [ ] Forgejo PR [Aditya]: Open PR from `feature/m3-skill-schema` to `master` with description "Add skill YAML schema, examples, loader, and tests.".
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git branch -d feature/m3-skill-schema`
- [ ] **COMMIT (Owner: Jeff | Group: C0.skill.domain | Branch: feature/m3-skill-domain) - Commit message: "feat(skill): add skill domain model and resolver"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-skill-domain`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Add `Skill`, `SkillItem`, `SkillToolRef`, `SkillInclude`, and `SkillInlineTool` models in `src/cleveragents/domain/models/core/skill.py` with namespaced naming rules.
  - [ ] Code [Jeff]: Implement `SkillResolver` to flatten includes into ordered tool lists, de-duplicate tools, and reject cycles with path traces.
  - [ ] Code [Jeff]: Add `Skill.resolve_tools()` returning resolved tool/validation names plus inline tool definitions for compiler use.
  - [ ] Docs [Jeff]: Add `docs/reference/skill_model.md` and `docs/reference/skill_resolution.md` with resolution order examples.
  - [ ] Tests (Behave) [Jeff]: Add `features/skill_resolution.feature` for include ordering, de-dupe rules, and cycle detection.
  - [ ] Tests (Robot) [Jeff]: Add `robot/skill_resolution.robot` smoke tests for resolver output.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/skill_resolution_bench.py` for resolver performance.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(skill): add skill domain model and resolver"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-skill-domain` to `master` with description "Add skill domain model, resolver, and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-skill-domain`
- [ ] **COMMIT (Owner: Luis | Group: C0.skill.registry | Branch: feature/m3-skill-registry) - Commit message: "feat(skill): add skill registry persistence"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m3-skill-registry`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Add `skills` and `skill_items` tables (namespaced name PK, description, source, yaml_text, timestamps) with indexes on namespace/name.
  - [ ] Code [Luis]: Implement `SkillRepository` CRUD + list filters and `SkillRegistryService` with add/update/remove/show/list.
  - [ ] Code [Luis]: Enforce referential integrity for included skills and tool references at registration time.
  - [ ] Docs [Luis]: Add `docs/reference/skill_registry.md` with registration and update behavior.
  - [ ] Tests (Behave) [Luis]: Add `features/skill_registry.feature` for add/update/remove and invalid include cases.
  - [ ] Tests (Robot) [Luis]: Add `robot/skill_registry.robot` CLI/service smoke tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/skill_registry_bench.py` for registry list performance.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(skill): add skill registry persistence"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m3-skill-registry` to `master` with description "Add skill registry persistence and service with tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m3-skill-registry`
- [ ] **COMMIT (Owner: Aditya | Group: C0.skill.cli | Branch: feature/m3-skill-cli) - Commit message: "feat(cli): add skill commands"**
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git pull origin master`
  - [ ] Git [Aditya]: `git checkout -b feature/m3-skill-cli`
  - [ ] Git [Aditya]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Aditya]: Implement `agents skill add/remove/list/show/tools` with YAML config input, schema validation, and `--namespace` filter.
  - [ ] Code [Aditya]: Add `--update` to `skill add` and emit clear errors for name mismatches.
  - [ ] Code [Aditya]: Ensure `skill tools` shows resolved tool list, inline tool IDs, validation nodes, and capability summaries.
  - [ ] Code [Aditya]: Add `--format json/yaml` outputs for list/show/tools with stable field ordering.
  - [ ] Docs [Aditya]: Update CLI reference with skill commands, examples, and output fields.
  - [ ] Tests (Behave) [Aditya]: Add CLI scenarios for skill add/show/tools/list/remove and invalid include cycles.
  - [ ] Tests (Robot) [Aditya]: Add `robot/skill_cli.robot` for end-to-end CLI flows.
  - [ ] Tests (ASV) [Aditya]: Add `asv/benchmarks/skill_cli_bench.py` for config parsing overhead.
  - [ ] Quality [Aditya]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Aditya]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Aditya]: `git add .` (only after coverage check passes)
  - [ ] Commit [Aditya]: `git commit -m "feat(cli): add skill commands"`.
  - [ ] Forgejo PR [Aditya]: Open PR from `feature/m3-skill-cli` to `master` with description "Add skill CLI commands, formats, and tests.".
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git branch -d feature/m3-skill-cli`

**Parallel Group C0.runtime: Tool Lifecycle Runtime [Jeff]** (depends on C0.domain + C0.registry; must land before C3.context)
- [ ] **COMMIT (Owner: Jeff | Group: C0.runtime | Branch: feature/m3-tool-runtime) - Commit message: "feat(tool): add tool lifecycle runtime"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-tool-runtime`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Implement `ToolRuntime`/`ToolInstance` interfaces with `discover/activate/execute/deactivate` hooks and lifecycle state tracking.
  - [ ] Code [Jeff]: Add `ToolExecutionContext` with resolved resource bindings, sandbox paths, plan metadata, and cancellation token.
  - [ ] Code [Jeff]: Add lifecycle cache with per-plan activation reuse and guaranteed `deactivate` on plan completion/cancel.
  - [ ] Code [Jeff]: Enforce tool capability flags (read-only/writes/checkpointable) and read-only plan gating at runtime.
  - [ ] Code [Jeff]: Validate tool inputs/outputs against JSON schema before/after execution; surface schema errors clearly.
  - [ ] Code [Jeff]: Add tool execution tracing (start/end timestamps, duration, result size) for diagnostics.
  - [ ] Code [Jeff]: Add cancellation propagation so long-running tools are interrupted on plan cancel.
  - [ ] Docs [Jeff]: Add `docs/reference/tool_lifecycle.md` describing hook ordering and failure handling.
  - [ ] Docs [Jeff]: Document schema validation behavior and error payload format for tool failures.
  - [ ] Tests (Behave) [Jeff]: Add lifecycle scenarios for activate/execute/deactivate ordering and error propagation.
  - [ ] Tests (Robot) [Jeff]: Add `robot/tool_lifecycle.robot` runtime smoke tests.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/tool_lifecycle_bench.py` for lifecycle overhead.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(tool): add tool lifecycle runtime"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-tool-runtime` to `master` with description "Add tool lifecycle runtime, context, and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-tool-runtime`

**Parallel Group C1: Actor Schema & Examples [Aditya + Jeff]** (start Day 5; C2 depends on this)
- [ ] **COMMIT (Owner: Aditya | Group: C1.schema | Branch: feature/m3-actor-schema-examples) - Commit message: "feat(actor): add actor yaml schema models"**
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git pull origin master`
  - [ ] Git [Aditya]: `git checkout -b feature/m3-actor-schema-examples`
  - [ ] Git [Aditya]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Aditya]: Add schema models (ActorType, NodeType, ContextView, ToolDefinition, RouteDefinition, ActorConfigSchema) with strict validation in `src/cleveragents/actor/schema.py`.
  - [ ] Code [Aditya]: Add tool-node schema fields that reference Tool Registry names, including validation nodes, and require input/output schema presence.
  - [ ] Code [Aditya]: Add hierarchical graph node schema (`children`, `edges`, `entrypoint`, `exitpoints`) to support nested actors and subgraphs.
  - [ ] Code [Aditya]: Add YAML load/serialize helpers and schema version guard.
  - [ ] Code [Aditya]: Add graph validation rules (entrypoint exists, exitpoints reachable, no orphan nodes).
  - [ ] Code [Aditya]: Add cycle detection for actor graphs and explicit errors for invalid loops.
  - [ ] Docs [Aditya]: Add `docs/reference/actors_schema.md` with field definitions, tool node semantics, and graph constraints.
  - [ ] Docs [Aditya]: Add graph validation examples (valid/invalid) and error messages.
  - [ ] Tests (Behave) [Aditya]: Add `features/actor_schema.feature` scenarios for validation and topology errors.
  - [ ] Tests (Robot) [Aditya]: Add `robot/actor_schema.robot` YAML load smoke test.
  - [ ] Tests (ASV) [Aditya]: Add `asv/benchmarks/actor_schema_bench.py` for YAML validation cost.
  - [ ] Quality [Aditya]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Aditya]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Aditya]: `git add .` (only after coverage check passes)
  - [ ] Commit [Aditya]: `git commit -m "feat(actor): add actor yaml schema models"`.
  - [ ] Forgejo PR [Aditya]: Open PR from `feature/m3-actor-schema-examples` to `master` with description "Add actor YAML schema models, validation rules, and tests.".
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git branch -d feature/m3-actor-schema-examples`
- [ ] **COMMIT (Owner: Aditya | Group: C1.examples | Branch: feature/m3-actor-schema-examples) - Commit message: "docs(actor): add actor yaml examples"**
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git pull origin master`
  - [ ] Git [Aditya]: `git checkout -b feature/m3-actor-schema-examples`
  - [ ] Git [Aditya]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Docs [Aditya]: Add `docs/reference/actors_examples.md` with strategist, executor, reviewer, tool-only, validation-node, and graph YAML examples.
  - [ ] Docs [Aditya]: Include hierarchical actor graph examples (graph-with-subgraph, multi-level planner/executor) aligned with spec.
  - [ ] Docs [Aditya]: Store example YAML files under `examples/actors/` for automated tests.
  - [ ] Tests (Behave) [Aditya]: Add `features/actor_examples.feature` to ensure all examples validate.
  - [ ] Tests (Robot) [Aditya]: Add `robot/actor_examples.robot` to load each example.
  - [ ] Tests (ASV) [Aditya]: Add `asv/benchmarks/actor_examples_load_bench.py` for YAML load throughput.
  - [ ] Quality [Aditya]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Aditya]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Aditya]: `git add .` (only after coverage check passes)
  - [ ] Commit [Aditya]: `git commit -m "docs(actor): add actor yaml examples"`.
  - [ ] Forgejo PR [Aditya]: Open PR from `feature/m3-actor-schema-examples` to `master` with description "Add actor YAML examples and validation tests.".
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git branch -d feature/m3-actor-schema-examples`

**Parallel Group C2: Actor Loading & Compilation [Aditya + Jeff]** (depends on C1)
  **PARALLEL SUBTRACK C2.legacy [Jeff]**: Remove v2 actor config compatibility (after C1.schema)
  **SEQUENTIAL NOTE**: C2.legacy must land before C2.loader/C2.compiler to avoid dual-format support.
- [ ] **COMMIT (Owner: Aditya | Group: C2.loader | Branch: feature/m3-actor-loader) - Commit message: "feat(actor): add actor registry and loader"**
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git pull origin master`
  - [ ] Git [Aditya]: `git checkout -b feature/m3-actor-loader`
  - [ ] Git [Aditya]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Aditya]: Implement actor loader/registry with namespaced lookup, cache invalidation, and file discovery in `actors/` and `examples/actors/`.
  - [ ] Code [Aditya]: Normalize actor file paths, reject non-`.yaml` files, and emit a single consolidated error when duplicates are found across search roots.
  - [ ] Code [Aditya]: Store and compare a content hash for actor configs to avoid reloading unchanged actors (mtime + hash fallback).
  - [ ] Code [Aditya]: Add registry integration with Tool Registry so tool nodes resolve at load time.
  - [ ] Code [Aditya]: Validate actor namespaced names on load and apply default `local/` when namespace omitted.
  - [ ] Docs [Aditya]: Add `docs/reference/actors_loading.md` with discovery rules and namespaces.
  - [ ] Tests (Behave) [Aditya]: Add `features/actor_loading.feature` for discovery, duplicates, and namespace lookup.
  - [ ] Tests (Robot) [Aditya]: Add `robot/actor_loading.robot` for loader smoke tests.
  - [ ] Tests (ASV) [Aditya]: Add `asv/benchmarks/actor_loading_bench.py` for registry load performance.
  - [ ] Quality [Aditya]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Aditya]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Aditya]: `git add .` (only after coverage check passes)
  - [ ] Commit [Aditya]: `git commit -m "feat(actor): add actor registry and loader"`.
  - [ ] Forgejo PR [Aditya]: Open PR from `feature/m3-actor-loader` to `master` with description "Add actor registry/loader, discovery rules, and tests.".
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git branch -d feature/m3-actor-loader`
- [ ] **COMMIT (Owner: Jeff | Group: C2.compiler | Branch: feature/m3-actor-compiler) - Commit message: "feat(actor): compile actor configs to LangGraph"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-actor-compiler`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Implement ActorCompiler that builds LangGraph for LLM, TOOL, and GRAPH actors with tool node wiring.
  - [ ] Code [Jeff]: Resolve tool node references through Tool Registry and validate required bindings before compile.
  - [ ] Code [Jeff]: Support validation nodes as first-class tool nodes and ensure they are marked read-only in compiled graphs.
  - [ ] Code [Jeff]: Enforce deterministic node ordering in compiled graphs for stable test snapshots.
  - [ ] Code [Jeff]: Emit compile-time diagnostics that include actor name, node id, and field path on error.
  - [ ] Docs [Jeff]: Add `docs/reference/actors_compilation.md` covering compile outputs and error modes.
  - [ ] Tests (Behave) [Jeff]: Add `features/actor_compilation.feature` for LLM/GRAPH compilation and tool node wiring.
  - [ ] Tests (Robot) [Jeff]: Add `robot/actor_compilation.robot` smoke test compiling all examples.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/actor_compilation_bench.py` for compilation speed.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(actor): compile actor configs to LangGraph"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-actor-compiler` to `master` with description "Compile actor configs to LangGraph with tool wiring and diagnostics.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-actor-compiler`
- [ ] **COMMIT (Owner: Jeff | Group: C2.refs | Branch: feature/m3-actor-refs) - Commit message: "feat(actor): resolve actor references and subgraphs"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-actor-refs`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Implement reference resolution, cycle detection, and subgraph wiring for actor refs.
  - [ ] Code [Jeff]: Ensure cross-namespace reference resolution follows `[server:]namespace/name` rules.
  - [ ] Code [Jeff]: Add explicit errors for ambiguous refs (same name in multiple registries) and include candidate list in error message.
  - [ ] Code [Jeff]: Preserve parent/child graph boundaries and annotate compiled graphs with source actor names for debugging.
  - [ ] Docs [Jeff]: Update `docs/reference/actors_compilation.md` with reference semantics.
  - [ ] Tests (Behave) [Jeff]: Add `features/actor_reference_resolution.feature` for missing/recursive refs.
  - [ ] Tests (Robot) [Jeff]: Add `robot/actor_reference_resolution.robot` for subgraph wiring.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/actor_reference_bench.py` for reference resolution performance.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(actor): resolve actor references and subgraphs"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-actor-refs` to `master` with description "Resolve actor references/subgraphs with diagnostics and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-actor-refs`

- [ ] **COMMIT (Owner: Jeff | Group: C2.legacy | Branch: feature/m3-actor-legacy-cleanup) - Commit message: "refactor(actor): drop v2 actor config compatibility"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-actor-legacy-cleanup`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Remove v2 JSON/YAML parsing paths in `src/cleveragents/actor/config.py` and related template engine usage.
  - [ ] Code [Jeff]: Ensure only v3 actor YAML schema is accepted; provide clear error message when v2 fields are present.
  - [ ] Code [Jeff]: Remove legacy actor config fixtures from `examples/` and update any docs that reference v2 file layouts.
  - [ ] Docs [Jeff]: Update `docs/reference/actors_loading.md` with v3-only note and migration guidance.
  - [ ] Tests (Behave) [Jeff]: Add scenarios that reject v2 actor config files.
  - [ ] Tests (Robot) [Jeff]: Add Robot tests that attempt to load v2 configs and assert failure.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/actor_schema_reject_bench.py` for validation overhead.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "refactor(actor): drop v2 actor config compatibility"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-actor-legacy-cleanup` to `master` with description "Drop v2 actor config compatibility and update docs/tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-actor-legacy-cleanup`

**Parallel Group C2.context: Actor Context Commands [Jeff]** (depends on C2.loader; post-M2)
- [ ] **COMMIT (Owner: Jeff | Group: C2.context | Branch: feature/m3-actor-context) - Commit message: "feat(cli): add actor context commands"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-actor-context`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Implement `agents actor context list/show/export/import/clear/remove` with `--all` and name filters per spec.
  - [ ] Code [Jeff]: Ensure actor context storage uses configured data dir and supports JSON import/export format.
  - [ ] Code [Jeff]: Add guardrails for missing contexts and unsafe overwrite confirmations.
  - [ ] Code [Jeff]: Add stable sorting for `context list` and `context show` output fields (consistent across formats).
  - [ ] Code [Jeff]: Enforce max context file size and return explicit error when exceeded.
  - [ ] Docs [Jeff]: Update CLI reference with actor context examples and file format notes.
  - [ ] Tests (Behave) [Jeff]: Add `features/actor_context_cli.feature` covering list/show/export/import/clear/remove.
  - [ ] Tests (Robot) [Jeff]: Add `robot/actor_context_cli.robot` for context command smoke tests.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/actor_context_cli_bench.py` for command overhead.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(cli): add actor context commands"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-actor-context` to `master` with description "Add actor context CLI commands and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-actor-context`

**Parallel Group C3: Skill Protocol & Context [Jeff]** (critical path; depends on C1)
- [ ] **COMMIT (Owner: Jeff | Group: C3.protocol | Branch: feature/m3-skill-protocol) - Commit message: "feat(skill): add skill protocol and metadata"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-skill-protocol`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Define Skill protocol interface, SkillMetadata, SkillResult, and SkillError types in `src/cleveragents/skills/protocol.py`.
  - [ ] Code [Jeff]: Add `SkillDefinition` model that references Tool Registry names and optional inline tool definitions.
  - [ ] Code [Jeff]: Add error mapping helpers to normalize tool failures into SkillError payloads.
  - [ ] Code [Jeff]: Require explicit `writes`/`read_only` metadata on skills and propagate to ToolRuntime gating.
  - [ ] Code [Jeff]: Add schema validation for `inputs`/`outputs` in SkillDefinition to match Tool Registry schemas.
  - [ ] Docs [Jeff]: Add `docs/reference/skills_protocol.md` describing metadata, tool composition, and JSON schema rules.
  - [ ] Tests (Behave) [Jeff]: Add `features/skill_protocol.feature` for metadata validation and error capture.
  - [ ] Tests (Robot) [Jeff]: Add `robot/skill_protocol.robot` smoke tests.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/skill_protocol_bench.py` for validation throughput.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(skill): add skill protocol and metadata"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-skill-protocol` to `master` with description "Add skill protocol, metadata models, and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-skill-protocol`
- [ ] **COMMIT (Owner: Jeff | Group: C3.context | Branch: feature/m3-skill-protocol) - Commit message: "feat(skill): add skill context and registry"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-skill-protocol`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Implement SkillContext (plan/resource access, sandbox path, change tracker) and SkillRegistry in `src/cleveragents/skills/context.py`.
  - [ ] Code [Jeff]: Wire SkillRegistry to Tool Registry for tool resolution and validation node inclusion.
  - [ ] Code [Jeff]: Add context helpers for resolving bound resources and exposing plan metadata.
  - [ ] Code [Jeff]: Include change tracking hooks so every skill execution registers a ToolInvocation in ChangeSet.
  - [ ] Code [Jeff]: Ensure SkillContext enforces read-only plan restrictions (block write tools when plan is read-only).
  - [ ] Docs [Jeff]: Add `docs/reference/skills_context.md` with context fields and helper methods.
  - [ ] Tests (Behave) [Jeff]: Add `features/skill_context.feature` for sandboxed access and registry resolution.
  - [ ] Tests (Robot) [Jeff]: Add `robot/skill_context.robot` for registry smoke tests.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/skill_context_bench.py` for registry resolution overhead.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(skill): add skill context and registry"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-skill-protocol` to `master` with description "Add skill context, registry, and change tracking hooks.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-skill-protocol`
- [ ] **COMMIT (Owner: Jeff | Group: C3.inline | Branch: feature/m3-skill-protocol) - Commit message: "feat(skill): add inline tool executor"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-skill-protocol`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Implement inline tool execution with timeouts and restricted environment in `src/cleveragents/skills/inline_executor.py`.
  - [ ] Code [Jeff]: Ensure inline tools conform to Tool Registry schema and return structured results.
  - [ ] Code [Jeff]: Add safeguards for file/network access inside inline tools (local-only for MVP).
  - [ ] Code [Jeff]: Add max runtime and max output size guards; return structured error on timeout/overflow.
  - [ ] Docs [Jeff]: Add `docs/reference/skills_inline.md` with safety constraints.
  - [ ] Tests (Behave) [Jeff]: Add `features/skill_inline.feature` for execution and timeout handling.
  - [ ] Tests (Robot) [Jeff]: Add `robot/skill_inline.robot` for inline tool smoke tests.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/inline_tool_bench.py` for execution overhead.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(skill): add inline tool executor"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-skill-protocol` to `master` with description "Add inline tool executor with safety constraints and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-skill-protocol`

**Parallel Group C4: Built-in Skills [Jeff + Luis]** (depends on C3)
- [ ] **COMMIT (Owner: Jeff | Group: C4.file | Branch: feature/m3-skill-file-search) - Commit message: "feat(skill): add file operation skills"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-skill-file-search`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Implement ReadFile, WriteFile, EditFile, and DeleteFile tools with read_only enforcement.
  - [ ] Code [Jeff]: Register tools in Tool Registry with resource bindings for fs/git resources and sandbox path rewrite.
  - [ ] Code [Jeff]: Add content size limits and encoding normalization (UTF-8) for file tools.
  - [ ] Code [Jeff]: Add path traversal guards (reject `..` and absolute paths outside sandbox root) and include rejected path in error.
  - [ ] Code [Jeff]: Make Write/Edit operations atomic (write temp file then replace) to avoid partial writes on failures.
  - [ ] Code [Jeff]: Add newline normalization option (preserve or enforce LF) with default preserve.
  - [ ] Code [Jeff]: Add binary file detection for ReadFile and emit explicit error for binary content.
  - [ ] Docs [Jeff]: Add `docs/reference/skills_file.md` with examples and error cases.
  - [ ] Docs [Jeff]: Document file size limits and binary file behavior.
  - [ ] Tests (Behave) [Jeff]: Add `features/skill_file_ops.feature` for read/write/edit/delete flows.
  - [ ] Tests (Robot) [Jeff]: Add `robot/skill_file_ops.robot` for file ops integration.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/file_tool_bench.py` for read/write throughput.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(skill): add file operation skills"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-skill-file-search` to `master` with description "Add file operation skills with sandboxed guards and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-skill-file-search`
- [ ] **COMMIT (Owner: Jeff | Group: C4.search | Branch: feature/m3-skill-file-search) - Commit message: "feat(skill): add directory and search skills"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-skill-file-search`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Implement ListDir, Glob, and Grep tools with ignore patterns and size limits.
  - [ ] Code [Jeff]: Register tools in Tool Registry with resource bindings and sandbox awareness.
  - [ ] Code [Jeff]: Enforce include/exclude glob filters from project context policies.
  - [ ] Code [Jeff]: Skip binary files and cap per-file match counts to keep output deterministic.
  - [ ] Code [Jeff]: Add regex compile error handling for Grep with explicit error output.
  - [ ] Code [Jeff]: Add sorting for glob/list outputs to ensure deterministic ordering.
  - [ ] Docs [Jeff]: Add `docs/reference/skills_search.md` with examples.
  - [ ] Docs [Jeff]: Document grep regex behavior and binary file skipping.
  - [ ] Tests (Behave) [Jeff]: Add `features/skill_search.feature` for listing/globbing/searching.
  - [ ] Tests (Robot) [Jeff]: Add `robot/skill_search.robot` for search integration.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/search_tool_bench.py` for search performance.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(skill): add directory and search skills"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-skill-file-search` to `master` with description "Add directory/search skills with deterministic output and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-skill-file-search`
- [ ] **COMMIT (Owner: Luis | Group: C4.git | Branch: feature/m3-skill-git) - Commit message: "feat(skill): add git operation skills"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m3-skill-git`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Implement read-only git tools (status, diff, log, show) for sandboxed repos.
  - [ ] Code [Luis]: Register git tools in Tool Registry with read-only capability metadata.
  - [ ] Code [Luis]: Add path guards to ensure git tools only run inside sandbox root.
  - [ ] Code [Luis]: Normalize git command output (strip color, set `GIT_PAGER=cat`) for deterministic tests.
  - [ ] Code [Luis]: Add safe environment variables for git execution (disable prompts, set safe.directory).
  - [ ] Code [Luis]: Add error mapping for common git failures (not a repo, detached HEAD) with explicit messages.
  - [ ] Docs [Luis]: Add `docs/reference/skills_git.md` clarifying no destructive ops in MVP.
  - [ ] Docs [Luis]: Document environment variables and safety settings for git tool execution.
  - [ ] Tests (Behave) [Luis]: Add `features/skill_git.feature` for git tool outputs.
  - [ ] Tests (Robot) [Luis]: Add `robot/skill_git.robot` for git tool integration.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/git_tool_bench.py` for diff/log performance.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(skill): add git operation skills"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m3-skill-git` to `master` with description "Add read-only git operation skills with safety guards and tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m3-skill-git`

**Parallel Group C5: Tool Routing & Change Tracking [Luis + Jeff]** (depends on C3/C4)
- [ ] **COMMIT (Owner: Luis | Group: C5.model | Branch: feature/m3-change-model) - Commit message: "feat(change): add ChangeSet models and invocation tracker"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m3-change-model`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Add Change/ChangeSet/ToolInvocation models and SkillInvocationTracker.
  - [ ] Code [Luis]: Ensure ChangeSet stores resource references, sandbox paths, tool metadata, and timestamps.
  - [ ] Code [Luis]: Add ChangeSet serialization helper for plan diff output (group by resource).
  - [ ] Code [Luis]: Enforce deterministic ordering for ChangeSet entries (by resource, then path, then timestamp).
  - [ ] Code [Luis]: Link ToolInvocation to plan_id and skill/tool names for auditability.
  - [ ] Code [Luis]: Add ChangeType enum (create/modify/delete/rename) and validate per-change required fields.
  - [ ] Code [Luis]: Add content hash fields (before_hash/after_hash) and file mode in Change for integrity.
  - [ ] Code [Luis]: Normalize file paths to repo-relative paths in ChangeSet entries.
  - [ ] Docs [Luis]: Add `docs/reference/change_tracking.md` describing tool-to-change mapping.
  - [ ] Tests (Behave) [Luis]: Add `features/change_tracking.feature` for ChangeSet aggregation.
  - [ ] Tests (Robot) [Luis]: Add `robot/change_tracking.robot` for tracker smoke tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/change_tracking_bench.py` for invocation tracking overhead.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(change): add ChangeSet models and invocation tracker"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m3-change-model` to `master` with description "Add ChangeSet models, invocation tracking, and tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m3-change-model`
- [ ] **COMMIT (Owner: Jeff | Group: C5.router | Branch: feature/m3-tool-router) - Commit message: "feat(change): add tool router for providers"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-tool-router`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Implement ToolCallRouter for OpenAI/Anthropic/LangChain tool schemas with deterministic IDs.
  - [ ] Code [Jeff]: Add mapping for tool/validation names and argument schemas based on Tool Registry metadata.
  - [ ] Code [Jeff]: Add tool-call result normalization to match ToolInvocation schema.
  - [ ] Code [Jeff]: Add provider-specific handling for streaming tool calls (accumulate args + finalize when complete).
  - [ ] Code [Jeff]: Ensure validation tools are surfaced with `tool_type=validation` in provider payloads.
  - [ ] Code [Jeff]: Normalize tool schemas to provider limits (field pruning, max description length) with explicit warnings.
  - [ ] Code [Jeff]: Add stable tool-call ID generation (plan_id + tool_name + sequence) and include in ToolInvocation metadata.
  - [ ] Code [Jeff]: Add error mapping for provider tool-call failures (timeout, schema error, tool not found) into ToolInvocation.error.
  - [ ] Code [Jeff]: Add provider metadata capture (model name, provider id, latency) for each tool call.
  - [ ] Docs [Jeff]: Add `docs/reference/tool_router.md` with provider-specific mappings.
  - [ ] Tests (Behave) [Jeff]: Add `features/tool_router.feature` for schema mapping.
  - [ ] Tests (Robot) [Jeff]: Add `robot/tool_router.robot` for routing smoke tests.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/tool_router_bench.py` for routing performance.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(change): add tool router for providers"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-tool-router` to `master` with description "Add tool router for providers with schema mapping and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-tool-router`
- [ ] **COMMIT (Owner: Luis | Group: C5.diff | Branch: feature/m3-diff-review) - Commit message: "feat(change): add diff review artifacts"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m3-diff-review`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Implement DiffBuilder and ReviewArtifact models for CLI review.
  - [ ] Code [Luis]: Add support for multi-resource diffs and per-resource grouping.
  - [ ] Code [Luis]: Add diff output serializers for rich/plain/json formats.
  - [ ] Code [Luis]: Include before/after content hashes and file modes in diff metadata for integrity checks.
  - [ ] Code [Luis]: Detect binary files and include a binary marker instead of inline diff content.
  - [ ] Code [Luis]: Add rename/move detection using path similarity + content hash and render as rename events.
  - [ ] Code [Luis]: Add max diff size guard with truncation notes and a `truncated=true` flag in JSON output.
  - [ ] Code [Luis]: Ensure diff output ordering is stable (resource name, path, change type).
  - [ ] Docs [Luis]: Add `docs/reference/diff_review.md` with output format.
  - [ ] Tests (Behave) [Luis]: Add `features/diff_review.feature` for diff generation.
  - [ ] Tests (Robot) [Luis]: Add `robot/diff_review.robot` for review artifacts.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/diff_review_bench.py` for diff building performance.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(change): add diff review artifacts"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m3-diff-review` to `master` with description "Add diff review artifacts and CLI-ready serializers with tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m3-diff-review`

**Parallel Group C6: Validation Pipeline [Luis + Jeff]** (depends on C5 and validation attachment config)
- [ ] **COMMIT (Owner: Luis | Group: C6.pipeline | Branch: feature/m3-validation-pipeline) - Commit message: "feat(validation): add validation pipeline and results model"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m3-validation-pipeline`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Implement ValidationCommand, ValidationResult, and ValidationPipeline using Validation attachments from Tool Registry.
  - [ ] Code [Luis]: Run validations at end of Execute phase only; do not re-run during Apply per spec.
  - [ ] Code [Luis]: Enforce required vs informational validation modes and fix-then-revalidate loop hooks.
  - [ ] Code [Luis]: Persist validation summary into Plan metadata for later review.
  - [ ] Code [Luis]: Sort validations by (resource, mode, validation name) for deterministic execution order.
  - [ ] Code [Luis]: Add per-validation timeout and capture stdout/stderr in ValidationResult for troubleshooting.
  - [ ] Code [Luis]: Group validations by resource and cap parallel validation concurrency per resource type.
  - [ ] Code [Luis]: Normalize validation output to `passed/message/data` schema even for tool errors.
  - [ ] Code [Luis]: Add aggregation helper to compute counts (required_passed, required_failed, info_failed) for CLI display.
  - [ ] Code [Luis]: Add guard to skip validations on read-only resources when validation requires write access (should never happen).
  - [ ] Docs [Luis]: Add `docs/reference/validation_pipeline.md` with ordering, timeouts, and failure handling.
  - [ ] Docs [Luis]: Document concurrency limits and output normalization rules.
  - [ ] Tests (Behave) [Luis]: Add `features/validation_pipeline.feature` for pass/fail paths and required/informational modes.
  - [ ] Tests (Robot) [Luis]: Add `robot/validation_pipeline.robot` for pipeline smoke tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/validation_pipeline_bench.py` for pipeline runtime.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(validation): add validation pipeline and results model"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m3-validation-pipeline` to `master` with description "Add validation pipeline and results model with ordering/timeout rules and tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m3-validation-pipeline`
- [ ] **COMMIT (Owner: Jeff | Group: C6.wraps | Branch: feature/m3-validation-wraps) - Commit message: "feat(validation): support wrapped tools and transforms"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-validation-wraps`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Implement validation `wraps` execution path that runs the wrapped Tool and captures its output.
  - [ ] Code [Jeff]: Add transform engine that maps wrapped tool output into ValidationResult schema (must output `passed` boolean).
  - [ ] Code [Jeff]: Enforce read-only constraints for validations even when wrapping write-capable tools; block if violation.
  - [ ] Code [Jeff]: Validate transform output schema at runtime and surface schema errors with tool name + validation name.
  - [ ] Code [Jeff]: Add guardrail that prevents `wraps` from invoking MCP tools in MVP (local-only) unless explicitly allowed.
  - [ ] Code [Jeff]: Add transform sandboxing (no imports, no file/network access) with explicit allowlist of safe helpers.
  - [ ] Code [Jeff]: Add transform execution timeout and max output size guard.
  - [ ] Code [Jeff]: Persist wrapped tool raw output in ValidationResult.data when transform fails for debugging.
  - [ ] Docs [Jeff]: Update `docs/reference/validation_model.md` with `wraps` + `transform` examples and safety rules.
  - [ ] Docs [Jeff]: Add transform error handling examples (schema failure, timeout).
  - [ ] Tests (Behave) [Jeff]: Add wrapped-validation scenarios with transform success/fail paths.
  - [ ] Tests (Robot) [Jeff]: Add `robot/validation_wraps.robot` for wrapped validation end-to-end.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/validation_wraps_bench.py` for transform overhead.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(validation): support wrapped tools and transforms"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-validation-wraps` to `master` with description "Support validation wraps/transform execution with safety guards and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-validation-wraps`
- [ ] **COMMIT (Owner: Jeff | Group: C6.gating | Branch: feature/m3-validation-gating) - Commit message: "feat(validation): integrate validation with apply gating"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-validation-gating`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Block apply on required validation failure; surface validation artifacts for review.
  - [ ] Code [Jeff]: Ensure informational validation failures do not block apply but are logged in plan status.
  - [ ] Code [Jeff]: Add CLI status output for validation summary (required vs informational counts).
  - [ ] Code [Jeff]: Ensure automation profile gates (manual/review/full-auto) influence whether apply can proceed automatically.
  - [ ] Code [Jeff]: Add explicit error when apply is attempted without Execute completion (phase/state guard).
  - [ ] Code [Jeff]: Add `plan status` fields for validation result IDs and last validation timestamp.
  - [ ] Code [Jeff]: Add apply gating hook that checks DoD evaluation before validations (fail fast).
  - [ ] Docs [Jeff]: Update `docs/reference/plan_actor_integration.md` with validation gating behavior.
  - [ ] Docs [Jeff]: Add gating flow diagram with manual/review/full-auto branches.
  - [ ] Tests (Behave) [Jeff]: Add `features/validation_gating.feature` for apply blocking.
  - [ ] Tests (Robot) [Jeff]: Add `robot/validation_gating.robot` for end-to-end gating.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/validation_gating_bench.py` for gating overhead.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(validation): integrate validation with apply gating"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-validation-gating` to `master` with description "Integrate validation gating into apply workflow with CLI summaries and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-validation-gating`

**Parallel Group C7: MCP Adapter [Aditya]** (depends on C3)
- [ ] **COMMIT (Owner: Aditya | Group: C7.mcp | Branch: feature/m3-mcp-adapter) - Commit message: "feat(skill): add MCP adapter for external tools"**
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git pull origin master`
  - [ ] Git [Aditya]: `git checkout -b feature/m3-mcp-adapter`
  - [ ] Git [Aditya]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Aditya]: Implement MCP client adapter conforming to Tool interface with connection config.
  - [ ] Code [Aditya]: Register MCP tools in Tool Registry with dynamic discovery from MCP server.
  - [ ] Code [Aditya]: Add timeout and retry defaults for MCP calls (local-only for MVP).
  - [ ] Code [Aditya]: Add connection health check and explicit error when MCP server is unreachable (include host/port).
  - [ ] Code [Aditya]: Cache MCP tool discovery results and expose a manual refresh command in the registry service.
  - [ ] Code [Aditya]: Add MCP server allowlist in config and block connections to non-allowlisted hosts.
  - [ ] Code [Aditya]: Add per-tool capability mapping (read/write/checkpointable) from MCP metadata where provided.
  - [ ] Code [Aditya]: Add local-only guard to prevent MCP execution in read-only plans unless tool is read-only.
  - [ ] Docs [Aditya]: Add `docs/reference/skills_mcp.md` with server connection examples.
  - [ ] Docs [Aditya]: Document allowlist config keys and refresh behavior.
  - [ ] Tests (Behave) [Aditya]: Add `features/skill_mcp.feature` for MCP tool calls.
  - [ ] Tests (Robot) [Aditya]: Add `robot/skill_mcp.robot` for MCP adapter smoke test.
  - [ ] Tests (ASV) [Aditya]: Add `asv/benchmarks/mcp_adapter_bench.py` for tool invocation latency.
  - [ ] Quality [Aditya]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Aditya]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Aditya]: `git add .` (only after coverage check passes)
  - [ ] Commit [Aditya]: `git commit -m "feat(skill): add MCP adapter for external tools"`.
  - [ ] Forgejo PR [Aditya]: Open PR from `feature/m3-mcp-adapter` to `master` with description "Add MCP adapter for external tools with allowlist and tests.".
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git branch -d feature/m3-mcp-adapter`

**Parallel Group C8: Built-in Provider Actors [Aditya]** (depends on C1/C2)
- [ ] **COMMIT (Owner: Aditya | Group: C8.providers | Branch: feature/m3-provider-actors) - Commit message: "feat(actor): add built-in provider actors"**
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git pull origin master`
  - [ ] Git [Aditya]: `git checkout -b feature/m3-provider-actors`
  - [ ] Git [Aditya]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Aditya]: Add built-in actor configs for `openai/`, `anthropic/`, and `openrouter/` (plus `google/` if configured).
  - [ ] Code [Aditya]: Add built-in actors for invariant reconciliation and estimation roles (using provider defaults).
  - [ ] Code [Aditya]: Set provider-specific defaults for temperature, max tokens, and tool-calling mode in each actor config.
  - [ ] Code [Aditya]: Add per-provider model mapping config keys and default model fallbacks.
  - [ ] Code [Aditya]: Add validation that provider actor configs reference existing provider adapters.
  - [ ] Code [Aditya]: Add example actor configs for local/llm-mock to support offline testing.
  - [ ] Docs [Aditya]: Add `docs/reference/provider_actors.md` with provider defaults.
  - [ ] Docs [Aditya]: Add environment variable mapping table for provider API keys and model defaults.
  - [ ] Tests (Behave) [Aditya]: Add `features/provider_actors.feature` for built-in actor loading.
  - [ ] Tests (Robot) [Aditya]: Add `robot/provider_actors.robot` for registry visibility.
  - [ ] Tests (ASV) [Aditya]: Add `asv/benchmarks/provider_actor_load_bench.py` for registry load cost.
  - [ ] Quality [Aditya]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Aditya]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Aditya]: `git add .` (only after coverage check passes)
  - [ ] Commit [Aditya]: `git commit -m "feat(actor): add built-in provider actors"`.
  - [ ] Forgejo PR [Aditya]: Open PR from `feature/m3-provider-actors` to `master` with description "Add built-in provider actor configs with defaults and tests.".
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git branch -d feature/m3-provider-actors`

**Parallel Group C9: Plan-Actor Integration [Jeff + Luis]** (depends on C2/C5/C6)
- [ ] **COMMIT (Owner: Jeff | Group: C9.execute | Branch: feature/m3-plan-execute) - Commit message: "feat(plan): execute strategize and execute phases via actors"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-plan-execute`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Connect PlanLifecycleService to actor execution for Strategize and Execute phases.
  - [ ] Code [Jeff]: Ensure Strategize is read-only and records decisions without modifying resources.
  - [ ] Code [Jeff]: Ensure Execute uses sandbox resources and tool calls routed through Tool Router + ChangeSet.
  - [ ] Code [Jeff]: Add plan status updates for phase start/complete/fail during actor execution.
  - [ ] Code [Jeff]: Persist decision tree root id and strategy decisions into Plan metadata after Strategize completes.
  - [ ] Code [Jeff]: Propagate project invariants + action invariants into the strategize actor context.
  - [ ] Code [Jeff]: Add explicit error handling for actor execution failures (capture error_message + error_details).
  - [ ] Code [Jeff]: Add execution metadata capture (tool_calls count, changeset_id, sandbox_refs) into Plan.
  - [ ] Code [Jeff]: Add streaming support hooks for `--stream` to emit interim status updates.
  - [ ] Code [Jeff]: Add guard to prevent Execute when Plan is not in Strategize COMPLETE state.
  - [ ] Docs [Jeff]: Add `docs/reference/plan_actor_integration.md` with phase flow.
  - [ ] Docs [Jeff]: Add error handling and retry notes for Strategize/Execute phases.
  - [ ] Tests (Behave) [Jeff]: Add `features/plan_actor_integration.feature` for strategy/execute flows.
  - [ ] Tests (Robot) [Jeff]: Add `robot/plan_actor_integration.robot` for end-to-end actor execution.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/plan_actor_integration_bench.py` for execution overhead.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(plan): execute strategize and execute phases via actors"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-plan-execute` to `master` with description "Integrate actor execution for strategize/execute phases with change tracking.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-plan-execute`
- [ ] **COMMIT (Owner: Jeff | Group: C9.apply | Branch: feature/m3-plan-apply) - Commit message: "feat(plan): integrate change review and apply flow"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m3-plan-apply`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Wire ChangeSet review artifacts into `plan diff` and review-before-apply flow.
  - [ ] Code [Jeff]: Ensure Apply merges sandbox into real resources only after required validations pass.
  - [ ] Code [Jeff]: Persist apply summary (files changed, validations) back into Plan metadata for `plan status`.
  - [ ] Code [Jeff]: Record apply timestamps and set plan state to terminal `applied` only after merge succeeds.
  - [ ] Code [Jeff]: Add merge failure handling with rollback to sandbox and explicit error output.
  - [ ] Code [Jeff]: Add per-resource apply summaries and include in plan artifacts output.
  - [ ] Code [Jeff]: Add guard to prevent Apply when ChangeSet is empty unless `--allow-empty` is set.
  - [ ] Docs [Jeff]: Update CLI docs for `plan diff` and `plan apply` review output.
  - [ ] Docs [Jeff]: Add apply error cases and recovery steps (re-run execute, fix validation).
  - [ ] Tests (Behave) [Jeff]: Add `features/plan_review_apply.feature` for review gate behavior.
  - [ ] Tests (Robot) [Jeff]: Add `robot/plan_review_apply.robot` for review-before-apply path.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/plan_apply_bench.py` for apply throughput.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(plan): integrate change review and apply flow"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m3-plan-apply` to `master` with description "Integrate change review and apply flow with validation gating and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m3-plan-apply`

**M3 SUCCESS CRITERIA**:
- Actor YAML schema validated; examples load and compile to LangGraph.
- Skills execute via SkillContext; built-in file/dir/search/git skills available.
- Tool-based change tracking (no output parsing) produces ChangeSet and diff review artifacts.
- MCP adapter executes a tool against a test MCP server.
- Built-in provider actors available (`openai/`, `anthropic/`, `openrouter/` as configured).
- Validation pipeline runs validation attachments and blocks apply on required failure.
- Plan lifecycle uses actors for Strategize/Execute and applies ChangeSet after review.
- `nox` passes with coverage >=97% across actor/skill/change-tracking suites.

**--- MERGE POINT 1: After M3, all workstreams coordinate ---**

---

### Section 6: Execution Pipeline, Decisions & Invariants [M3-M4]

**Target: Milestone M4 (+21 days)**
**Week 3 focus**: decision capture, correction, invariants, and DoD gating.

**Parallel Group D1: Decision Domain [Hamza]** (foundation for D2-D5; tests supported by Brent when Rui is unavailable)
- [ ] **COMMIT (Owner: Hamza | Group: D1.domain | Branch: feature/m4-decision-domain) - Commit message: "feat(domain): add decision model and context snapshots"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m4-decision-domain`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add `DecisionType`, `ContextSnapshot`, and `Decision` models with correction fields and helpers.
  - [ ] Code [Hamza]: Include required fields: question, chosen option, alternatives, confidence score, rationale, dependencies, and context hash.
  - [ ] Code [Hamza]: Add identifiers (`decision_id` ULID, `plan_id`, `parent_decision_id`, `root_decision_id`) and supersession markers (`superseded_by`, `superseded_at`).
  - [ ] Code [Hamza]: Add `ContextSnapshot` fields for `detail_level`, `token_count`, `source_refs`, and `created_at` with hash-based dedupe helper.
  - [ ] Code [Hamza]: Add validators to enforce `confidence_score` in 0.0–1.0 range and non-empty question text.
  - [ ] Code [Hamza]: Add `DecisionSummary` helper for compact CLI rendering (question, choice, confidence, decision_id).
  - [ ] Docs [Hamza]: Add `docs/reference/decision_model.md` with examples and schema notes.
  - [ ] Tests (Behave) [Hamza]: Add `features/decision_model.feature` for validation and helpers.
  - [ ] Tests (Robot) [Hamza]: Add `robot/decision_model.robot` smoke tests.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/decision_model_bench.py` for decision validation throughput.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(domain): add decision model and context snapshots"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m4-decision-domain` to `master` with description "Add decision domain models and context snapshot helpers.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m4-decision-domain`

**Parallel Group D2: Decision Recording Service [Hamza]** (depends on D1)
- [ ] **COMMIT (Owner: Hamza | Group: D2.service | Branch: feature/m4-decision-service) - Commit message: "feat(service): add decision recording and snapshot store"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m4-decision-service`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Implement `DecisionService` with `record_decision`, sequence numbers, tree queries, and downstream linking.
  - [ ] Code [Hamza]: Add `ContextSnapshotStore` interface with a file-backed MVP implementation and hash dedupe.
  - [ ] Code [Hamza]: Integrate decision recording into strategize/execute phases (prompt/strategy/subplan/tool decisions).
  - [ ] Code [Hamza]: Add `DecisionType` mapping for `prompt_definition`, `invariant_enforced`, `strategy_choice`, `subplan_spawn`, and `subplan_parallel_spawn`.
  - [ ] Code [Hamza]: Persist context snapshot references with each decision and store snapshots under data dir by hash.
  - [ ] Code [Hamza]: Add `DecisionService.record_context_snapshot()` helper to avoid duplicate snapshot writes.
  - [ ] Code [Hamza]: Add structured logging for decision recording (plan_id, decision_id, type, confidence).
  - [ ] Docs [Hamza]: Add `docs/reference/decision_service.md` covering recording and snapshots.
  - [ ] Tests (Behave) [Hamza]: Add `features/decision_recording.feature` scenarios.
  - [ ] Tests (Robot) [Hamza]: Add `robot/decision_recording.robot` integration smoke tests.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/decision_recording_bench.py` for record throughput.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(service): add decision recording and snapshot store"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m4-decision-service` to `master` with description "Add decision recording service and snapshot store with tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m4-decision-service`

**Parallel Group D3: Decision CLI & Viewing [Hamza]** (depends on D1/D2)
- [ ] **COMMIT (Owner: Hamza | Group: D3.cli | Branch: feature/m4-decision-cli) - Commit message: "feat(cli): add plan tree and explain commands"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m4-decision-cli`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Implement `plan tree` and `plan explain` with rich/json/flat formats and `--show-superseded`/`--show-context`.
  - [ ] Code [Hamza]: Add `--show-reasoning` to include confidence and alternatives in explain output per spec.
  - [ ] Code [Hamza]: Add `--depth` and `--type` filters for tree output to reduce large plan noise.
  - [ ] Code [Hamza]: Add deterministic tree ordering (by sequence number, then decision_id) to stabilize snapshots.
  - [ ] Code [Hamza]: Add `plan explain --format json/yaml` output schema with decision fields and context snapshot refs.
  - [ ] Docs [Hamza]: Update CLI reference for decision viewing commands.
  - [ ] Docs [Hamza]: Include output field mapping table for `plan tree` and `plan explain`.
  - [ ] Tests (Behave) [Hamza]: Add tree/explain scenarios including superseded handling.
  - [ ] Tests (Robot) [Hamza]: Add `robot/decision_cli.robot` smoke tests.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/decision_cli_bench.py` for tree rendering overhead.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(cli): add plan tree and explain commands"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m4-decision-cli` to `master` with description "Add plan tree/explain commands and decision viewing outputs.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m4-decision-cli`

**Parallel Group D4: Decision Correction [Jeff + Luis]** (depends on D2/D3)
- [ ] **COMMIT (Owner: Jeff | Group: D4.revert | Branch: feature/m4-decision-revert) - Commit message: "feat(service): add decision correction revert flow"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m4-decision-revert`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Implement correction impact analysis and dry-run reporting.
  - [ ] Code [Jeff]: Revert flow with checkpoint rollback, supersede downstream decisions, and subtree re-exec.
  - [ ] Code [Jeff]: Persist correction attempt IDs and link them to superseded decisions.
  - [ ] Code [Jeff]: Record correction attempts with user guidance text and timestamp for audit trail.
  - [ ] Code [Jeff]: Re-run validations after re-exec and update plan status with correction outcome summary.
  - [ ] Code [Jeff]: Add `plan correct --dry-run` output schema (affected decisions, affected subplans, estimated re-exec count).
  - [ ] Code [Jeff]: Add correction state machine (pending -> running -> applied/failed) and store in plan metadata.
  - [ ] Code [Jeff]: Ensure corrections cannot run when plan is in APPLY/APPLIED (explicit error message).
  - [ ] Docs [Jeff]: Add `docs/reference/decision_correction.md` for revert behavior.
  - [ ] Docs [Jeff]: Include dry-run output example and correction state transitions.
  - [ ] Tests (Behave) [Jeff]: Add revert + dry-run scenarios.
  - [ ] Tests (Robot) [Jeff]: Add revert integration tests with checkpoint rollback.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/decision_correction_revert_bench.py` for correction overhead.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(service): add decision correction revert flow"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m4-decision-revert` to `master` with description "Add decision correction revert flow with dry-run and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m4-decision-revert`
- [ ] **COMMIT (Owner: Jeff | Group: D4.append | Branch: feature/m4-decision-append) - Commit message: "feat(service): add decision correction append flow"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m4-decision-append`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Append flow creating fix subplan without rewriting history; link correction attempt + decision tree updates.
  - [ ] Code [Jeff]: Record append corrections as separate subtree with explicit lineage.
  - [ ] Code [Jeff]: Enforce that append corrections do not modify prior decisions; they only add a new branch and mark prior branch as superseded.
  - [ ] Code [Jeff]: Add plan metadata linking correction attempt to spawned subplan IDs.
  - [ ] Code [Jeff]: Add `plan correct --mode=append` dry-run output with proposed subplan count.
  - [ ] Docs [Jeff]: Extend correction docs for append mode and guidance-file usage.
  - [ ] Docs [Jeff]: Add sample append correction flow with resulting decision tree snippet.
  - [ ] Tests (Behave) [Jeff]: Add append correction scenarios.
  - [ ] Tests (Robot) [Jeff]: Add append correction smoke test.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/decision_correction_append_bench.py` for append overhead.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(service): add decision correction append flow"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m4-decision-append` to `master` with description "Add decision correction append flow and docs/tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m4-decision-append`

**Parallel Group D5: Decision Persistence [Hamza + Luis]** (depends on D1)
- [ ] **COMMIT (Owner: Hamza | Group: D5.db | Branch: feature/m4-decision-persistence) - Commit message: "feat(db): add decision tables"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m4-decision-persistence`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add Alembic migrations for `decisions` and `context_snapshots` with indexes.
  - [ ] Code [Hamza]: Add indexes for plan_id, decision_type, and superseded flags for fast tree queries.
  - [ ] Code [Hamza]: Store decision JSON payloads with deterministic ordering for stable diffs and audits.
  - [ ] Docs [Hamza]: Update `docs/reference/database_schema.md` with decision tables.
  - [ ] Tests (Behave) [Hamza]: Add migration verification scenarios.
  - [ ] Tests (Robot) [Hamza]: Add DB migration smoke test.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/decision_migration_bench.py` for migration baseline.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(db): add decision tables"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m4-decision-persistence` to `master` with description "Add decision persistence tables and migrations.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m4-decision-persistence`
- [ ] **COMMIT (Owner: Hamza | Group: D5.repo | Branch: feature/m4-decision-persistence) - Commit message: "feat(repo): add decision repositories"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m4-decision-persistence`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Implement DecisionRepository + ContextSnapshotRepository with tree queries and max-sequence helpers.
  - [ ] Code [Hamza]: Add repository methods for superseded decision lookup and subtree retrieval.
  - [ ] Code [Hamza]: Add repository method to fetch ordered decision path from root to a given decision (for explain output).
  - [ ] Docs [Hamza]: Document repository interfaces in `docs/reference/repositories.md`.
  - [ ] Tests (Behave) [Hamza]: Add decision persistence scenarios (create/query/superseded).
  - [ ] Tests (Robot) [Hamza]: Add repository integration smoke test.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/decision_repository_bench.py` for tree query performance.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(repo): add decision repositories"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m4-decision-persistence` to `master` with description "Add decision repositories and persistence queries.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m4-decision-persistence`
- [ ] **COMMIT (Owner: Luis | Group: D5.di | Branch: feature/m4-decision-di) - Commit message: "feat(di): wire decision services"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m4-decision-di`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Wire decision repositories + services into DI and CLI.
  - [ ] Code [Luis]: Inject DecisionService into PlanLifecycleService so strategize/execute can record decisions automatically.
  - [ ] Docs [Luis]: Update DI docs for decision wiring.
  - [ ] Tests (Behave) [Luis]: Add DI wiring scenarios for decision commands.
  - [ ] Tests (Robot) [Luis]: Add CLI smoke test using persisted decisions.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/decision_di_bench.py` for DI resolution overhead.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(di): wire decision services"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m4-decision-di` to `master` with description "Wire decision services into DI and CLI with tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m4-decision-di`
- [ ] **COMMIT (Owner: Brent | Group: D5.tests | Branch: feature/m4-decision-tests) - Commit message: "test(persistence): add decision persistence suites"**
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git pull origin master`
  - [ ] Git [Brent]: `git checkout -b feature/m4-decision-tests`
  - [ ] Git [Brent]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Tests (Behave) [Brent]: Add `features/decision_persistence.feature` scenarios.
  - [ ] Tests (Behave) [Brent]: Add scenarios for superseded decisions, correction attempts, and tree depth filtering.
  - [ ] Tests (Behave) [Brent]: Add output snapshots for `plan explain --format json` and `plan tree --format yaml`.
  - [ ] Tests (Robot) [Brent]: Add `robot/decision_persistence.robot` E2E coverage.
  - [ ] Docs [Brent]: Update `docs/development/testing.md` with decision suites.
  - [ ] Tests (ASV) [Brent]: Add `asv/benchmarks/decision_persistence_bench.py` for DB persistence throughput.
  - [ ] Quality [Brent]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Brent]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Brent]: `git add .` (only after coverage check passes)
  - [ ] Commit [Brent]: `git commit -m "test(persistence): add decision persistence suites"`.
  - [ ] Forgejo PR [Brent]: Open PR from `feature/m4-decision-tests` to `master` with description "Add decision persistence test suites and benchmarks.".
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git branch -d feature/m4-decision-tests`

**Parallel Group DOD: Definition of Done + Invariants [Luis + Jeff]** (depends on D2/D4)
- [ ] **COMMIT (Owner: Luis | Group: DOD.dod | Branch: feature/m4-definition-of-done) - Commit message: "feat(dod): enforce definition-of-done gating"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m4-definition-of-done`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Evaluate `definition_of_done` before apply; block apply with clear error if unmet.
  - [ ] Code [Luis]: Ensure DoD templating uses plan arguments and preserves template in plan metadata.
  - [ ] Code [Luis]: Add DoD evaluation hook to capture pass/fail reasoning and store in plan validation summary.
  - [ ] Code [Luis]: Add DoD evaluator interface with a default text matcher implementation (regex/substring checks).
  - [ ] Code [Luis]: Add CLI output in `plan status` showing DoD evaluation summary (passed/failed + message).
  - [ ] Docs [Luis]: Add `docs/reference/definition_of_done.md` with examples.
  - [ ] Docs [Luis]: Include a template rendering example with plan args and a failure output example.
  - [ ] Tests (Behave) [Luis]: Add DoD pass/fail scenarios.
  - [ ] Tests (Robot) [Luis]: Add DoD integration smoke test.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/dod_evaluation_bench.py` for evaluation overhead.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(dod): enforce definition-of-done gating"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m4-definition-of-done` to `master` with description "Enforce definition-of-done gating with evaluation and tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m4-definition-of-done`
- [ ] **COMMIT (Owner: Jeff | Group: DOD.invariants | Branch: feature/m4-invariants) - Commit message: "feat(invariant): add invariant models and enforcement"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m4-invariants`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Add invariant models, merge order (plan > project > action > global), and enforcement before strategize.
  - [ ] Code [Jeff]: Add Invariant Reconciliation Actor role and record `invariant_enforced` decisions.
  - [ ] Code [Jeff]: Add `agents invariant add/list/remove` CLI with scope flags.
  - [ ] Code [Jeff]: Persist invariants in DB with scope fields and integrate with plan creation defaults.
  - [ ] Code [Jeff]: Add explicit error when invariant text is empty or duplicates an existing invariant in the same scope.
  - [ ] Code [Jeff]: Add `InvariantScope` enum (global/project/action/plan) and normalize scope precedence for merges.
  - [ ] Code [Jeff]: Add CLI output showing invariant source scope and attachment ID for removal.
  - [ ] Code [Jeff]: Add migration for `invariants` table and `invariant_attachments` table (project/plan/action references).
  - [ ] Docs [Jeff]: Add `docs/reference/invariants.md` and update CLI reference.
  - [ ] Docs [Jeff]: Add examples for scope precedence and reconciliation actor outputs.
  - [ ] Tests (Behave) [Jeff]: Add invariant merge + violation scenarios.
  - [ ] Tests (Robot) [Jeff]: Add invariant CLI integration tests.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/invariant_merge_bench.py` for merge overhead.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(invariant): add invariant models and enforcement"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m4-invariants` to `master` with description "Add invariant models, persistence, and enforcement with tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m4-invariants`


### Section 7: Subplans & Parallelism [M5]

**Target: Milestone M5 (+25 days)**
**Week 3-4 focus**: subplan spawning, parallel execution, and result merging.

**Parallel Group E1: Subplan Domain [Luis]** (tests supported by Brent while Rui is unavailable)
- [ ] **COMMIT (Owner: Luis | Group: E1.domain | Branch: feature/m5-subplan-domain) - Commit message: "feat(domain): add subplan config and status models"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m5-subplan-domain`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Add `ExecutionMode` enum (sequential, parallel, hybrid) with validation guards.
  - [ ] Code [Luis]: Add `MergeStrategy` enum (three_way, sequential, json) with defaults.
  - [ ] Code [Luis]: Add `SubplanConfig` model fields: parent_plan_id, spawn_decision_id, dependencies, max_parallel, merge_strategy, automation_profile_override, invariants_override, context_view_override.
  - [ ] Code [Luis]: Add `SubplanStatus` model fields: subplan_id, state, started_at, completed_at, error_message, changeset_id.
  - [ ] Code [Luis]: Add `SubplanAttempt` model fields: attempt_id (ULID), subplan_id, attempt_number, started_at, completed_at, error_details.
  - [ ] Code [Luis]: Add `SubplanState` enum (queued/processing/errored/complete/cancelled) and validators to align with ProcessingState semantics.
  - [ ] Code [Luis]: Add `SubplanConfig` validation for `max_parallel >= 1`, `merge_strategy` compatibility, and non-empty dependency IDs.
  - [ ] Code [Luis]: Add `SubplanConfig.to_decision_payload()`/`from_decision_payload()` helpers for strategy actor handoff.
  - [ ] Code [Luis]: Add `Plan.link_child_subplan(child_id, decision_id)` helper to keep deterministic child ordering.
  - [ ] Code [Luis]: Add serialization ordering rules so subplan configs render with stable field ordering (for snapshot tests).
  - [ ] Code [Luis]: Extend `Plan` with `subplan_config`, `subplan_statuses`, `spawn_decision_id`, and helpers (`is_subplan`, `has_subplans`, `child_count`).
  - [ ] Code [Luis]: Add DecisionType constants for `subplan_spawn` and `subplan_parallel_spawn` and ensure models reference them.
  - [ ] Code [Luis]: Add dependency validation helper to reject cycles and missing dependency references.
  - [ ] Docs [Luis]: Add `docs/reference/subplan_model.md`.
  - [ ] Docs [Luis]: Include field-by-field table with defaults + example JSON payloads for decision emission.
  - [ ] Tests (Behave) [Luis]: Add `features/subplan_model.feature` scenarios for config validation, dependency cycles, and parent/root helpers.
  - [ ] Tests (Robot) [Luis]: Add `robot/subplan_model.robot` smoke tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/subplan_model_bench.py` for model validation.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(domain): add subplan config and status models"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m5-subplan-domain` to `master` with description "Add subplan domain models, enums, and tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m5-subplan-domain`

- [X] **Stage E1: Subplan Model** (Day 12) **[Luis]**

  **SEQUENTIAL ORDER**: E1.1 (Enums) → E1.2 (SubplanConfig) → E1.3 (Plan extension) → E1.4 (SubplanStatus) → E1.5 (Failure rules) → E1.6 (Tests)

  - [X] **E1.1** [Luis] Define execution enums in `src/cleveragents/domain/models/core/plan.py`:
    - [X] **E1.1a** [Luis] Define `ExecutionMode` enum:
      ```python
      class ExecutionMode(str, Enum):
          """How subplans should be executed."""
          SEQUENTIAL = "sequential"       # One after another, ordered by sequence
          PARALLEL = "parallel"           # All at once (up to max_parallel)
          DEPENDENCY_ORDERED = "dependency_ordered"  # Respect DAG dependencies
      ```
      - [X] Commit: "feat(domain): define ExecutionMode enum"
    - [X] **E1.1b** [Luis] Define `SubplanMergeStrategy` enum (renamed from `MergeStrategy` to avoid collision with infrastructure merge):
      ```python
      class SubplanMergeStrategy(str, Enum):
          """How to merge results from parallel subplans."""
          GIT_THREE_WAY = "git_three_way"  # Use git merge-file for code
          SEQUENTIAL_APPLY = "sequential_apply"  # Apply in completion order
          FAIL_ON_CONFLICT = "fail_on_conflict"  # Error if any conflicts
          LAST_WINS = "last_wins"  # Later changes overwrite earlier
      ```
      - [X] Commit: "feat(domain): define SubplanMergeStrategy enum"
  - [X] **E1.2** [Luis] Define `SubplanConfig` model:
    - [X] **E1.2a** [Luis] Create SubplanConfig dataclass:
      ```python
      class SubplanConfig(BaseModel):
          """Configuration for subplan execution."""

          execution_mode: ExecutionMode = Field(
              default=ExecutionMode.SEQUENTIAL,
              description="How to execute subplans"
          )
          merge_strategy: SubplanMergeStrategy = Field(
              default=SubplanMergeStrategy.GIT_THREE_WAY,
              description="How to merge subplan results"
          )
          max_parallel: int = Field(
              default=5, ge=1, le=50,
              description="Max concurrent subplans (for PARALLEL mode)"
          )
          fail_fast: bool = Field(
              default=False,
              description="Stop all subplans on first failure"
          )
          timeout_per_subplan_seconds: int | None = Field(
              default=None,
              description="Timeout for each subplan (None=no timeout)"
          )
          retry_failed: bool = Field(
              default=True,
              description="Automatically retry failed subplans"
          )
          max_retries: int = Field(
              default=2, ge=0, le=5,
              description="Max retry attempts per subplan"
          )
      ```
      - [X] Commit: "feat(domain): define SubplanConfig model"
  - [X] **E1.3** [Luis] Extend Plan model for subplan hierarchy:
    - [X] **E1.3a** [Luis] Add parent/root plan fields (verify exist):
      ```python
      # In PlanIdentity model (already existed from A1)
      parent_plan_id: str | None = Field(
          default=None,
          description="Parent plan ID if this is a subplan"
      )
      root_plan_id: str | None = Field(
          default=None,
          description="Root plan ID (topmost ancestor)"
      )
      ```
      - [X] Commit: "feat(domain): verify parent/root plan fields on Plan"
    - [X] **E1.3b** [Luis] Add subplan configuration field:
      ```python
      subplan_config: SubplanConfig | None = Field(
          default=None,
          description="Config for subplan execution (set on parent plans)"
      )
      subplan_statuses: list["SubplanStatus"] = Field(
          default_factory=list,
          description="Status tracking for spawned subplans"
      )
      ```
      - [X] Commit: "feat(domain): add subplan config and status fields to Plan"
    - [X] **E1.3c** [Luis] Add computed properties:
      ```python
      @property
      def is_subplan(self) -> bool:
          """Check if this plan is a subplan (has parent)."""
          return self.parent_plan_id is not None

      @property
      def is_root_plan(self) -> bool:
          """Check if this is the root plan."""
          return self.root_plan_id is None or self.root_plan_id == self.plan_id

      @property
      def depth(self) -> int:
          """Distance from root plan (0 for root)."""
          # Note: This requires parent chain traversal
          # For efficiency, may be cached or stored
          if self.is_root_plan:
              return 0
          # Computed by service layer traversing parent_plan_id chain
          return -1  # Placeholder, computed externally

      @property
      def has_subplans(self) -> bool:
          """Check if this plan has spawned subplans."""
          return len(self.subplan_statuses) > 0
      ```
      - [X] Commit: "feat(domain): add subplan computed properties to Plan"
  - [X] **E1.4** [Luis] Define `SubplanStatus` tracking model:
    - [X] **E1.4a** [Luis] Create SubplanStatus dataclass:
      ```python
      @dataclass
      class SubplanStatus:
          """Track status of a spawned subplan."""

          subplan_id: str  # The subplan's plan_id
          action_name: str  # Action used to create subplan
          target_resources: list[str]  # Resources subplan works on

          # Status tracking
          status: ProcessingState = ProcessingState.QUEUED
          started_at: datetime | None = None
          completed_at: datetime | None = None

          # Results
          error: str | None = None
          changeset_summary: str | None = None  # Brief summary of changes
          files_changed: int = 0

          # Retries
          attempt_number: int = 1
          previous_attempts: list["SubplanAttempt"] = field(default_factory=list)
      ```
      - [X] Commit: "feat(domain): define SubplanStatus dataclass"
    - [X] **E1.4b** [Luis] Define SubplanAttempt for retry tracking:
      ```python
      @dataclass
      class SubplanAttempt:
          """Record of a subplan execution attempt."""
          attempt_number: int
          started_at: datetime
          completed_at: datetime | None
          error: str | None
          was_retried: bool
      ```
      - [X] Commit: "feat(domain): define SubplanAttempt dataclass"
  - [X] **E1.5** [Luis] Define subplan failure handling rules:
    - [X] **E1.5a** [Luis] Create `SubplanFailureHandler` class:
      ```python
      class SubplanFailureHandler:
          """Handle subplan failures based on configuration."""

          def should_stop_others(
              self,
              config: SubplanConfig,
              failed_status: SubplanStatus
          ) -> bool:
              """Determine if other subplans should stop."""
              if config.fail_fast:
                  return True
              if config.execution_mode == ExecutionMode.SEQUENTIAL:
                  return True  # Sequential always stops on failure
              return False  # Parallel continues others

          def should_retry(
              self,
              config: SubplanConfig,
              status: SubplanStatus
          ) -> bool:
              """Determine if failed subplan should be retried."""
              if not config.retry_failed:
                  return False
              if status.attempt_number > config.max_retries:
                  return False
              # Don't retry on certain errors
              if status.error and "ValidationError" in status.error:
                  return True  # Validation failures can be retried
              if status.error and "TimeoutError" in status.error:
                  return True  # Timeouts can be retried
              return False
      ```
      - [X] Commit: "feat(domain): define SubplanFailureHandler"
    - [X] **E1.5b** [Luis] Add failure state constants:
      ```python
      # Error = application/system bug, likely not recoverable
      # Failure = task couldn't complete (tests fail, validation fail), may be retryable

      RETRIABLE_FAILURES = {
          "ValidationError",
          "TimeoutError",
          "TemporaryResourceError",
          "MergeConflictError"  # May succeed with different merge strategy
      }

      NON_RETRIABLE_ERRORS = {
          "ConfigurationError",
          "AuthenticationError",
          "MissingResourceError",
          "CircularDependencyError"
      }
      ```
      - [X] Commit: "feat(domain): define retriable vs non-retriable failures"
  - [ ] **E1.6** [Brent] Write Behave tests for subplan model:
    - [ ] **E1.6a** [Brent] Plan hierarchy scenarios:
      - [ ] Scenario: Plan with parent_plan_id has is_subplan=True
        - [ ] Given Plan with parent_plan_id set
        - [ ] Then is_subplan returns True
        - [ ] And is_root_plan returns False
      - [ ] Scenario: Root plan has is_subplan=False and is_root_plan=True
      - [ ] Scenario: SubplanConfig validates max_parallel bounds
      - [ ] Commit: "test(behave): add plan hierarchy scenarios"
    - [ ] **E1.6b** [Brent] Execution mode scenarios:
      - [ ] Scenario: ExecutionMode enum has all required values
      - [ ] Scenario: MergeStrategy enum has all required values
      - [ ] Scenario: SubplanConfig defaults are applied
      - [ ] Commit: "test(behave): add execution mode scenarios"
    - [ ] **E1.6c** [Brent] SubplanStatus scenarios:
      - [ ] Scenario: SubplanStatus tracks state correctly
      - [ ] Scenario: SubplanAttempt records retry history
      - [ ] Scenario: Failure handler respects fail_fast setting
      - [ ] Commit: "test(behave): add SubplanStatus scenarios"

**Parallel Group E2: Subplan Spawning [Jeff + Aditya]** (depends on D2 + E1)
- [ ] **COMMIT (Owner: Jeff | Group: E2.service | Branch: feature/m5-subplan-service) - Commit message: "feat(service): add subplan service and spawn workflow"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m5-subplan-service`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Implement `SubplanService` with `spawn_subplan`, `spawn_batch`, tree queries, and bounded context builder.
  - [ ] Code [Jeff]: Build bounded context from parent plan decisions + project context policies; enforce token/file limits.
  - [ ] Code [Jeff]: Inherit automation profile + invariants from parent plan; allow subplan overrides from decisions.
  - [ ] Code [Jeff]: Persist subplan config into child Plan metadata (`subplan_config`) and link `spawn_decision_id`.
  - [ ] Code [Jeff]: Link SUBPLAN_SPAWN decisions to created subplans and status tracking.
  - [ ] Code [Jeff]: Enforce `max_parallel` and dependency ordering at spawn time; reject invalid dependency lists with explicit error.
  - [ ] Code [Jeff]: Add idempotency guard to avoid double-spawn on retries (hash plan_id + decision_id).
  - [ ] Code [Jeff]: Add `SubplanService.list_children(parent_plan_id)` and `get_child_statuses(parent_plan_id)` helpers for CLI/status output.
  - [ ] Code [Jeff]: Add spawn audit log entry with decision_id, child_plan_id, and derived context stats.
  - [ ] Code [Jeff]: Normalize subplan names to include parent namespaced name + sequence number for readability.
  - [ ] Code [Jeff]: Add explicit error when requested resource scope is not linked to parent project(s).
  - [ ] Docs [Jeff]: Add `docs/reference/subplan_service.md`.
  - [ ] Docs [Jeff]: Document idempotency, naming convention, and spawn audit log format.
  - [ ] Tests (Behave) [Jeff]: Add subplan spawn scenarios (inheritance, overrides, dependency ordering).
  - [ ] Tests (Robot) [Jeff]: Add subplan spawn integration tests.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/subplan_spawn_bench.py` for spawn throughput.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(service): add subplan service and spawn workflow"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m5-subplan-service` to `master` with description "Add subplan service, spawn workflows, and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m5-subplan-service`
- [ ] **COMMIT (Owner: Aditya | Group: E2.actor | Branch: feature/m5-subplan-actor) - Commit message: "feat(actor): add plan_subplan tool and decision emission"**
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git pull origin master`
  - [ ] Git [Aditya]: `git checkout -b feature/m5-subplan-actor`
  - [ ] Git [Aditya]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Aditya]: Add `plan_subplan` tool to strategy actors and emit SUBPLAN_SPAWN decisions.
  - [ ] Code [Aditya]: Support `parallel=true` to emit SUBPLAN_PARALLEL_SPAWN and include dependency list.
  - [ ] Code [Aditya]: Include merge strategy, resource scope, and context view overrides in decision payload.
  - [ ] Code [Aditya]: Validate that subplan tool payload includes at least one resource scope or project ref.
  - [ ] Code [Aditya]: Add schema validation for `max_parallel`, `merge_strategy`, and dependency list in tool payload.
  - [ ] Code [Aditya]: Add defaulting rules for omitted fields (inherit from action/plan; explicit override markers).
  - [ ] Code [Aditya]: Emit decision rationale text with summary of why subplan was spawned (for explain output).
  - [ ] Docs [Aditya]: Update actor YAML examples for subplan emission.
  - [ ] Docs [Aditya]: Add a minimal `plan_subplan` tool payload example and a parallel spawn example with dependencies.
  - [ ] Tests (Behave) [Aditya]: Add scenarios for subplan decision emission (parallel + dependencies).
  - [ ] Tests (Robot) [Aditya]: Add actor tool integration smoke tests.
  - [ ] Tests (ASV) [Aditya]: Add `asv/benchmarks/subplan_actor_tool_bench.py` for tool invocation overhead.
  - [ ] Quality [Aditya]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Aditya]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Aditya]: `git add .` (only after coverage check passes)
  - [ ] Commit [Aditya]: `git commit -m "feat(actor): add plan_subplan tool and decision emission"`.
  - [ ] Forgejo PR [Aditya]: Open PR from `feature/m5-subplan-actor` to `master` with description "Add plan_subplan tool emission rules with tests and examples.".
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git branch -d feature/m5-subplan-actor`

**Parallel Group E3: Parallel Execution [Luis + Jeff]** (depends on E1/E2)
- [ ] **COMMIT (Owner: Luis | Group: E3.exec | Branch: feature/m5-subplan-exec) - Commit message: "feat(service): add subplan scheduler and execution"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m5-subplan-exec`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Add subplan scheduler with `max_parallel`, dependency ordering, and fail-fast handling.
  - [ ] Code [Luis]: Support sequential and parallel execution modes based on SubplanConfig.
  - [ ] Code [Luis]: Track status updates for subplans and propagate to parent plan (processing/complete/errored).
  - [ ] Code [Luis]: Add cancellation propagation from parent to child subplans.
  - [ ] Code [Luis]: Add worker-pool abstraction with bounded concurrency and deterministic scheduling order.
  - [ ] Code [Luis]: Add retry/backoff policy for failed subplan execution (configurable per automation profile).
  - [ ] Code [Luis]: Persist execution timestamps for each subplan attempt and expose in parent plan status.
  - [ ] Code [Luis]: Add scheduler telemetry counters (queued/running/completed/errored) for diagnostics.
  - [ ] Code [Luis]: Ensure scheduler honors read-only plans (reject write-capable tools in subplans).
  - [ ] Docs [Luis]: Add `docs/reference/subplan_execution.md`.
  - [ ] Docs [Luis]: Document scheduler ordering, retry policy, and cancellation propagation.
  - [ ] Tests (Behave) [Luis]: Add parallel + dependency execution scenarios.
  - [ ] Tests (Robot) [Luis]: Add parallel execution integration tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/subplan_scheduler_bench.py` for scheduler overhead.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(service): add subplan scheduler and execution"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m5-subplan-exec` to `master` with description "Add subplan scheduler and execution with dependency ordering and tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m5-subplan-exec`

**Parallel Group E4: Result Merging [Jeff]** (depends on E3)
- [ ] **COMMIT (Owner: Jeff | Group: E4.merge | Branch: feature/m5-subplan-merge) - Commit message: "feat(merge): add subplan merge strategies"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m5-subplan-merge`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Add three-way merge strategy for file changes and conflict markers.
  - [ ] Code [Jeff]: Add sequential merge and JSON merge strategies; expose merge result artifacts.
  - [ ] Code [Jeff]: Add conflict artifact model (file_path, conflict_type, base/left/right snippets).
  - [ ] Code [Jeff]: Store merge output as ChangeSet and attach to parent plan for review.
  - [ ] Code [Jeff]: Add merge summary metadata (files merged, conflicts count) for `plan status` output.
  - [ ] Code [Jeff]: Add deterministic merge ordering (by resource, path) to make diff outputs stable for tests.
  - [ ] Code [Jeff]: Add merge conflict resolution hints in artifacts (suggested resolution path).
  - [ ] Code [Jeff]: Add guard that blocks auto-apply when unresolved conflicts exist.
  - [ ] Docs [Jeff]: Add `docs/reference/subplan_merge.md`.
  - [ ] Docs [Jeff]: Include conflict artifact schema and sample output for `plan diff`.
  - [ ] Tests (Behave) [Jeff]: Add merge + conflict scenarios.
  - [ ] Tests (Robot) [Jeff]: Add merge integration tests for multi-subplan plans.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/subplan_merge_bench.py` for merge performance.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(merge): add subplan merge strategies"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m5-subplan-merge` to `master` with description "Add subplan merge strategies and conflict artifacts with tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m5-subplan-merge`

**Parallel Group E5: Multi-Project Plans [Hamza]** (depends on E2/E4)
- [ ] **COMMIT (Owner: Hamza | Group: E5.multi | Branch: feature/m5-multi-project) - Commit message: "feat(plan): add multi-project subplan support"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m5-multi-project`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Allow plans to target multiple projects with separate resource link contexts.
  - [ ] Code [Hamza]: Ensure sandbox isolation and cross-project dependency resolution.
  - [ ] Code [Hamza]: Add plan metadata to track project-specific ChangeSets and validation summaries.
  - [ ] Code [Hamza]: Add project-scoped context view resolution so each subplan only sees its project resources.
  - [ ] Code [Hamza]: Add `ProjectLink` alias resolution rules for multi-project plans (unique alias per plan).
  - [ ] Code [Hamza]: Add CLI output for `plan status` to show per-project ChangeSet summaries.
  - [ ] Code [Hamza]: Enforce that a plan cannot mix read-only and write-capable projects without explicit override.
  - [ ] Docs [Hamza]: Add `docs/reference/multi_project_plans.md`.
  - [ ] Docs [Hamza]: Add examples for multi-project `plan use` and explain alias resolution.
  - [ ] Tests (Behave) [Hamza]: Add multi-project subplan scenarios.
  - [ ] Tests (Robot) [Hamza]: Add multi-project integration tests.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/multi_project_bench.py` for multi-project overhead.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(plan): add multi-project subplan support"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m5-multi-project` to `master` with description "Add multi-project subplan support and per-project summaries.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m5-multi-project`


### Section 8: Large Project Autonomy & Context [M6]

**Target: Milestone M6 (+30 days)**
**Local-mode only**: large-project autonomy is required; server connectivity remains stubbed.

**Parallel Group G1: Large-Project Decomposition [Jeff]**
- [ ] **COMMIT (Owner: Jeff | Group: G1.decompose | Branch: feature/m6-large-decompose) - Commit message: "feat(plan): add large-project decomposition and dependency closure"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m6-large-decompose`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Add hierarchical decomposition with 4+ levels and bounded context per subplan.
  - [ ] Code [Jeff]: Implement decomposition heuristics (max_files_per_subplan, max_tokens_per_subplan, language/dir clustering).
  - [ ] Code [Jeff]: Add dependency closure computation for large graphs and DAG execution ordering.
  - [ ] Code [Jeff]: Add bounded dependency closure with cutoff thresholds and memoization for 10K+ files.
  - [ ] Code [Jeff]: Record decomposition decisions in DecisionService (strategy_choice + subplan_spawn entries).
  - [ ] Code [Jeff]: Add fallback decomposition strategy when heuristics fail (single subplan with warning).
  - [ ] Code [Jeff]: Add config keys in Settings for `planner.max_depth`, `planner.max_files_per_subplan`, `planner.max_tokens_per_subplan`, `planner.min_files_per_subplan` with defaults.
  - [ ] Code [Jeff]: Add deterministic clustering order (directory depth, language priority, file size) to make decomposition stable.
  - [ ] Code [Jeff]: Add decomposition summary metrics (subplan count, max depth, avg size) stored in plan metadata.
  - [ ] Code [Jeff]: Add guard to avoid spawning subplans when project is below threshold; default to single plan.
  - [ ] Docs [Jeff]: Add `docs/reference/large_project_decomposition.md`.
  - [ ] Docs [Jeff]: Document config keys, default thresholds, and example decomposition outputs.
  - [ ] Tests (Behave) [Jeff]: Add deep hierarchy + dependency closure scenarios.
  - [ ] Tests (Robot) [Jeff]: Add large-project decomposition integration tests.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/large_project_decompose_bench.py` for decomposition runtime.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(plan): add large-project decomposition and dependency closure"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m6-large-decompose` to `master` with description "Add large-project decomposition, dependency closure, and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m6-large-decompose`

**Parallel Group G2: Checkpointing & Rollback [Luis]**
- [ ] **COMMIT (Owner: Luis | Group: G2.checkpoint | Branch: feature/m6-checkpoint) - Commit message: "feat(checkpoint): add checkpointing and rollback"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m6-checkpoint`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Add checkpoint declarations for tools and plan-level rollback policy.
  - [ ] Code [Luis]: Add `checkpoints` table (checkpoint_id ULID, plan_id, sandbox_ref, created_at, metadata_json).
  - [ ] Code [Luis]: Implement `plan rollback <plan_id> <checkpoint_id>` command.
  - [ ] Code [Luis]: Implement git-worktree checkpoint snapshots (commit hash or patch) and rollback restore.
  - [ ] Code [Luis]: Wire rollback into decision correction revert flow for reuse.
  - [ ] Code [Luis]: Add checkpoint retention policy (max checkpoints per plan) with auto-prune on new checkpoints.
  - [ ] Code [Luis]: Store checkpoint reason/source (tool name, phase) in metadata for auditability.
  - [ ] Code [Luis]: Add CLI output to `plan rollback` showing restored file counts and changed paths.
  - [ ] Code [Luis]: Add guard preventing rollback when plan is applied or sandbox is missing.
  - [ ] Docs [Luis]: Add `docs/reference/checkpointing.md`.
  - [ ] Docs [Luis]: Include checkpoint retention defaults and rollback error cases.
  - [ ] Tests (Behave) [Luis]: Add checkpoint/rollback scenarios.
  - [ ] Tests (Robot) [Luis]: Add rollback integration tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/checkpoint_rollback_bench.py` for rollback latency.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(checkpoint): add checkpointing and rollback"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m6-checkpoint` to `master` with description "Add checkpointing and rollback support with tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m6-checkpoint`

**Parallel Group G3: Semantic Validation [Luis]**
- [ ] **COMMIT (Owner: Luis | Group: G3.semantic | Branch: feature/m6-semantic-validation) - Commit message: "feat(validation): add semantic validation service"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m6-semantic-validation`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Add semantic validation hooks during strategize/execute and error-pattern checks.
  - [ ] Code [Luis]: Add built-in semantic checks for syntax errors, missing imports, and broken references for Python projects.
  - [ ] Code [Luis]: Expose semantic validation as Validation tools so they can be attached per resource.
  - [ ] Code [Luis]: Add rule registry for semantic validators (dependency cycles, API misuse, missing symbols).
  - [ ] Code [Luis]: Integrate semantic validation results into ValidationPipeline as informational by default.
  - [ ] Code [Luis]: Add config keys for enabling/disabling semantic validation per project and per plan (default on for Python).
  - [ ] Code [Luis]: Add severity mapping (info/warn/error) and map required vs informational behavior.
  - [ ] Code [Luis]: Add output schema normalization so validations return `passed`, `message`, and `data` fields consistently.
  - [ ] Code [Luis]: Add caching for semantic checks keyed by file hash to avoid rework on unchanged files.
  - [ ] Docs [Luis]: Add `docs/reference/semantic_validation.md`.
  - [ ] Docs [Luis]: Add section on required vs informational validation attachment modes.
  - [ ] Tests (Behave) [Luis]: Add semantic validation scenarios.
  - [ ] Tests (Robot) [Luis]: Add semantic validation integration tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/semantic_validation_bench.py` for validation cost.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(validation): add semantic validation service"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m6-semantic-validation` to `master` with description "Add semantic validation service and attachments with tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m6-semantic-validation`

**Parallel Group G4: Context Tiers & Views [Hamza]**
- [ ] **COMMIT (Owner: Hamza | Group: G4.context | Branch: feature/m6-context-tiers) - Commit message: "feat(context): add hot/warm/cold tiers and actor views"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m6-context-tiers`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Implement hot/warm/cold tiers with indexing, LRU eviction, and promotion/demotion.
  - [ ] Code [Hamza]: Add tier storage backends (in-memory hot, sqlite warm, file-backed cold).
  - [ ] Code [Hamza]: Add per-actor context views (strategist/executor/reviewer) and filtered presentation.
  - [ ] Code [Hamza]: Add summarization hook when demoting to cold tier.
  - [ ] Code [Hamza]: Enforce project-scoped filtering (ScopedBackendView) so actors never see resources outside the plan's project set.
  - [ ] Code [Hamza]: Add `skeleton_ratio` support to cap inherited context size for subplans.
  - [ ] Code [Hamza]: Add tier budget settings (max_tokens_hot, max_decisions_warm, max_decisions_cold) and defaults per project.
  - [ ] Code [Hamza]: Add deterministic ordering for tier retrieval (most-recent-first with stable tie-breakers).
  - [ ] Code [Hamza]: Add cold-tier compaction job to merge adjacent summaries and cap file size.
  - [ ] Code [Hamza]: Add metrics for tier hit/miss counts and expose via `agents diagnostics`.
  - [ ] Docs [Hamza]: Add `docs/reference/context_tiers.md`.
  - [ ] Docs [Hamza]: Document tier budgets, eviction rules, and summarization policy.
  - [ ] Tests (Behave) [Hamza]: Add context tier scenarios.
  - [ ] Tests (Robot) [Hamza]: Add context tier integration tests.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/context_tiers_bench.py` for tier lookup performance.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(context): add hot/warm/cold tiers and actor views"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m6-context-tiers` to `master` with description "Add context tiers, actor views, and tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m6-context-tiers`

**Parallel Group G5: Cost & Risk Estimation [Hamza]**
- [ ] **COMMIT (Owner: Hamza | Group: G5.estimate | Branch: feature/m6-estimation) - Commit message: "feat(estimation): add cost and risk estimation actor"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m6-estimation`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add optional `estimation_actor` role and cost/risk estimation outputs.
  - [ ] Code [Hamza]: Persist estimation output to plan metadata (cost_estimate, risk_score, duration_estimate).
  - [ ] Code [Hamza]: Invoke estimation during `plan use` and surface estimates in `plan status` output.
  - [ ] Code [Hamza]: Add estimation output schema with currency, token estimates, and confidence ranges.
  - [ ] Code [Hamza]: Add opt-out flag `--no-estimate` for `plan use` and persist `estimate_skipped` reason.
  - [ ] Code [Hamza]: Add error handling when estimation actor fails (fallback to informational warning).
  - [ ] Docs [Hamza]: Add `docs/reference/estimation.md` with output format.
  - [ ] Docs [Hamza]: Add CLI examples showing estimate fields in `plan status` output.
  - [ ] Tests (Behave) [Hamza]: Add estimation scenarios.
  - [ ] Tests (Robot) [Hamza]: Add estimation integration smoke tests.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/estimation_actor_bench.py` for estimation runtime.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(estimation): add cost and risk estimation actor"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m6-estimation` to `master` with description "Add cost/risk estimation actor and plan outputs.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m6-estimation`

**Parallel Group G6: CLI Polish [Jeff]**
- [ ] **COMMIT (Owner: Jeff | Group: G6.cli | Branch: feature/m6-cli-polish) - Commit message: "chore(cli): polish help and output"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m6-cli-polish`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Standardize help text, progress indicators, and error messages with recovery hints.
  - [ ] Code [Jeff]: Ensure `--format` outputs are consistent (rich/color/table/plain/json/yaml) across core commands.
  - [ ] Code [Jeff]: Ensure `plain` format uses ASCII-only output to support log pipelines.
  - [ ] Code [Jeff]: Add stable column names and ordering for list commands (`action list`, `plan list`, `project list`, `resource list`).
  - [ ] Code [Jeff]: Add `--format json/yaml` schema docs and align field names across commands.
  - [ ] Code [Jeff]: Add unified error envelope for JSON/YAML outputs (`error.code`, `error.message`, `error.details`).
  - [ ] Docs [Jeff]: Update CLI output examples where needed.
  - [ ] Docs [Jeff]: Add a CLI output contract section in `docs/reference/cli_output.md`.
  - [ ] Tests (Behave) [Jeff]: Add `features/cli_output_formats.feature` covering rich/plain/json/yaml formatting for core commands.
  - [ ] Tests (Robot) [Jeff]: Add CLI UX smoke tests for critical commands.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/cli_render_bench.py` for output rendering overhead.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "chore(cli): polish help and output"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m6-cli-polish` to `master` with description "Polish CLI help/output formats and add tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m6-cli-polish`

**--- MERGE POINT 2: Day 30 - Large Project Autonomy Target (LOCAL MODE ONLY) ---**

By Day 30, the system must be able to (all in LOCAL MODE, server connectivity is deferred):
- Handle projects with 10,000+ files using hierarchical decomposition.
- Port source code from one language to another autonomously.
- Use hierarchical subplans (5+ levels deep) for large tasks.
- Correct decisions at any point without full re-execution.
- Operate entirely in local mode (server client stubs in place but not implemented).


### Section 9: Server Connectivity (Stubs Only) [By Day 30]

**Target: Deliver stubs by M6 (Day 30); full server implementation remains post-Day 30**

**Parallel Group F0: Server Client Stubs [Luis]** (required for M6; no server implementation)
- [ ] **COMMIT (Owner: Luis | Group: F0.stubs | Branch: feature/m6-server-stubs) - Commit message: "feat(interfaces): add server client stubs"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m6-server-stubs`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Add protocol stubs for `ServerClient`, `RemoteExecutionClient`, and `AuthClient` with NotImplementedError.
  - [ ] Code [Luis]: Add `ServerConnectionConfig` model (server_url, namespace, auth_token_ref, tls_verify) with validation.
  - [ ] Code [Luis]: Add config keys (`core.server_url`, `core.server_namespace`, `core.server_tls_verify`) and ensure they load from config/env.
  - [ ] Code [Luis]: Add `agents connect <server_url>` CLI stub in `cli/commands/server_client.py`.
  - [ ] Code [Luis]: Implement `agents connect` to persist server_url into config and print an explicit "server mode not implemented" warning.
  - [ ] Code [Luis]: Add `agents info` output field for Server Mode: `disabled` vs `stubbed` (based on config).
  - [ ] Code [Luis]: Add environment variable support (`CLEVERAGENTS_SERVER_URL`, `CLEVERAGENTS_SERVER_NAMESPACE`) with precedence documented.
  - [ ] Code [Luis]: Add stub methods that return explicit error envelopes in JSON/YAML outputs (avoid stack traces).
  - [ ] Code [Luis]: Add unit helper `is_server_enabled()` and use it to guard server-only commands.
  - [ ] Docs [Luis]: Add `docs/reference/server_client_stubs.md` noting client-only behavior.
  - [ ] Docs [Luis]: Document config precedence (CLI flag > config file > env) for server settings.
  - [ ] Tests (Behave) [Luis]: Add stub behavior scenarios.
  - [ ] Tests (Behave) [Luis]: Add scenarios for env var override and config precedence order.
  - [ ] Tests (Robot) [Luis]: Add CLI stub smoke test.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/server_stub_bench.py` (baseline no-op).
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(interfaces): add server client stubs"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m6-server-stubs` to `master` with description "Add server client stubs and config wiring with tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m6-server-stubs`

**M7 SUCCESS CRITERIA** (Post-Day 30):
- `agents [--data-dir PATH] [--config-path PATH] connect <server_url>` establishes connection to an external server.
- Plans can be synced and executed on a remote server.
- Real-time updates received via WebSocket from server.
- Remote projects can be specified and executed on server.


### Section 10: Async Infrastructure & Later-Stage Quality Work [Various Leads]

**Note**: Quality automation setup is in Section 0; Section 10 focuses on async infrastructure and later-stage validation support.

**Parallel Group 10A: Async Infrastructure [Luis]**
- [ ] **COMMIT (Owner: Luis | Group: 10A.async | Branch: feature/m6-async-infra) - Commit message: "feat(async): add async command execution and workers"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m6-async-infra`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Implement async command execution per ADR-002 with cancellation and timeout handling.
  - [ ] Code [Luis]: Add `AsyncJob` model and `async_jobs` table (plan_id, phase, status, payload_json, created_at, started_at, finished_at).
  - [ ] Code [Luis]: Add `AsyncJobStatus` enum and enforce valid transitions (queued -> running -> succeeded/failed/cancelled).
  - [ ] Code [Luis]: Add `worker_id` and `last_heartbeat` fields; mark stuck jobs failed after TTL.
  - [ ] Code [Luis]: Add AsyncWorker orchestrator with polling loop, max_workers config, and graceful shutdown hooks.
  - [ ] Code [Luis]: Add job enqueue hooks for plan execute/apply when async is enabled via config flag (no new CLI flags).
  - [ ] Code [Luis]: Add cancellation token support and ensure cancellation propagates to tool execution.
  - [ ] Code [Luis]: Add config keys `async.enabled`, `async.max_workers`, `async.poll_interval`, `async.job_timeout`, and `async.job_ttl`.
  - [ ] Code [Luis]: Add job cleanup routine to prune completed jobs older than a retention threshold.
  - [ ] Code [Luis]: Add worker health report (last_heartbeat, jobs processed) surfaced in `agents diagnostics`.
  - [ ] Code [Luis]: Add serialization of async payloads with schema version for forward compatibility.
  - [ ] Docs [Luis]: Update `docs/reference/async_architecture.md` with execution flow, job states, and shutdown rules.
  - [ ] Docs [Luis]: Document async config defaults and job retention policy.
  - [ ] Tests (Behave) [Luis]: Add `features/async_execution.feature` for async command handling (enqueue, worker pick-up, cancel).
  - [ ] Tests (Robot) [Luis]: Add `robot/async_execution.robot` smoke tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/async_execution_bench.py` for worker scheduling overhead.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(async): add async command execution and workers"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m6-async-infra` to `master` with description "Add async command execution, workers, and tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m6-async-infra`
- [ ] **COMMIT (Owner: Luis | Group: 10A.retry | Branch: feature/m6-async-infra) - Commit message: "feat(async): wire retry policies into services"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m6-async-infra`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Integrate retry/circuit breaker policies into service layer operations.
  - [ ] Code [Luis]: Add retry policy configuration keys (max_attempts, base_delay, max_delay, jitter) to settings.
  - [ ] Code [Luis]: Ensure retries are only applied to idempotent operations (repository reads, validation calls) and never to applies.
  - [ ] Code [Luis]: Add `RetryPolicy` and `CircuitBreaker` models with per-service overrides.
  - [ ] Code [Luis]: Emit structured logs for retry attempts and circuit-open events.
  - [ ] Code [Luis]: Add per-service policy defaults with overrides via config (service_name -> policy mapping).
  - [ ] Code [Luis]: Add guard to prevent retries on tool execution writes in read-only plans.
  - [ ] Code [Luis]: Add circuit breaker half-open recovery logic and cooldown timers.
  - [ ] Docs [Luis]: Document retry policy defaults and override points.
  - [ ] Docs [Luis]: Add examples of per-service override config and expected logs.
  - [ ] Tests (Behave) [Luis]: Add retry/circuit breaker behavior scenarios.
  - [ ] Tests (Robot) [Luis]: Add resilience smoke tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/retry_policy_bench.py` for retry overhead.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(async): wire retry policies into services"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m6-async-infra` to `master` with description "Wire retry/circuit breaker policies into services with tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m6-async-infra`

**Parallel Group 10B: Selective Quality Review [Brent]**
- [ ] **COMMIT (Owner: Brent | Group: 10B.review | Branch: feature/m6-review-playbook) - Commit message: "docs(qa): add review playbook and priority matrix"**
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git pull origin master`
  - [ ] Git [Brent]: `git checkout -b feature/m6-review-playbook`
  - [ ] Git [Brent]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Docs [Brent]: Create `docs/development/review_playbook.md` with focus areas and skip rules.
  - [ ] Docs [Brent]: Add priority matrix and review SLA guidance.
  - [ ] Docs [Brent]: Add checklist templates for architecture review, CLI review, and DB migration review.
  - [ ] Docs [Brent]: Add review checklists for security-sensitive changes (secrets, auth, server stubs) and schema migrations.
  - [ ] Docs [Brent]: Add a PR review routing table (which reviewer for which subsystem).
  - [ ] Docs [Brent]: Add examples of acceptable vs blocking findings with remediation guidance.
  - [ ] Docs [Brent]: Add a required test matrix section for reviewers (nox sessions + coverage).
  - [ ] Tests (Behave) [Brent]: Add scenarios validating review playbook references exist.
  - [ ] Tests (Robot) [Brent]: Add docs build smoke test covering the new guide.
  - [ ] Tests (ASV) [Brent]: Add `asv/benchmarks/docs_build_bench.py` for docs build baseline.
  - [ ] Quality [Brent]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Brent]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Brent]: `git add .` (only after coverage check passes)
  - [ ] Commit [Brent]: `git commit -m "docs(qa): add review playbook and priority matrix"`.
  - [ ] Forgejo PR [Brent]: Open PR from `feature/m6-review-playbook` to `master` with description "Add review playbook, priority matrix, and tests.".
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git branch -d feature/m6-review-playbook`

**Parallel Group 10C: Validation Testing Support [Brent + Luis]**
- [ ] **COMMIT (Owner: Brent | Group: 10C.edge | Branch: feature/m6-validation-edge) - Commit message: "test(validation): add edge case suites"**
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git pull origin master`
  - [ ] Git [Brent]: `git checkout -b feature/m6-validation-edge`
  - [ ] Git [Brent]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Brent]: Add shared edge-case fixtures under `features/fixtures/validation/`.
  - [ ] Code [Brent]: Add fixtures for malformed tool outputs, missing resources, and validation timeouts.
  - [ ] Code [Brent]: Add fixtures for wrapped validation transforms that return invalid schema.
  - [ ] Code [Brent]: Add fixtures for mixed required/informational validation ordering and duplicate attachment IDs.
  - [ ] Docs [Brent]: Update `docs/development/testing.md` with validation test catalog.
  - [ ] Tests (Behave) [Brent]: Add edge-case scenarios for concurrency, conflicts, rollbacks, and timeouts.
  - [ ] Tests (Robot) [Brent]: Add integration coverage for edge-case suites.
  - [ ] Tests (ASV) [Brent]: Add `asv/benchmarks/validation_edge_bench.py` for edge-case runtime.
  - [ ] Quality [Brent]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Brent]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Brent]: `git add .` (only after coverage check passes)
  - [ ] Commit [Brent]: `git commit -m "test(validation): add edge case suites"`.
  - [ ] Forgejo PR [Brent]: Open PR from `feature/m6-validation-edge` to `master` with description "Add validation edge-case suites and fixtures.".
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git branch -d feature/m6-validation-edge`
- [ ] **COMMIT (Owner: Luis | Group: 10C.semantic | Branch: feature/m6-validation-semantic) - Commit message: "test(validation): add semantic validation suites"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m6-validation-semantic`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Add semantic validation fixtures and error-pattern samples.
  - [ ] Code [Luis]: Add fixtures for language-porting mismatches and dependency graph violations.
  - [ ] Code [Luis]: Add fixtures for API surface changes (renamed functions, missing symbols, incompatible types).
  - [ ] Code [Luis]: Add fixtures for cross-file symbol resolution and circular import detection.
  - [ ] Docs [Luis]: Document semantic validation coverage expectations.
  - [ ] Tests (Behave) [Luis]: Add semantic validation scenarios.
  - [ ] Tests (Robot) [Luis]: Add semantic validation integration tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/semantic_validation_suite_bench.py` for suite runtime.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "test(validation): add semantic validation suites"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m6-validation-semantic` to `master` with description "Add semantic validation test suites and fixtures.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m6-validation-semantic`
- [ ] **COMMIT (Owner: Brent | Group: 10C.performance | Branch: feature/m6-perf-scale) - Commit message: "test(perf): add scale test fixtures"**
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git pull origin master`
  - [ ] Git [Brent]: `git checkout -b feature/m6-perf-scale`
  - [ ] Git [Brent]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Brent]: Add scale fixtures for 1K/5K/10K file repos in `features/fixtures/scale/`.
  - [ ] Code [Brent]: Add scriptless fixture generator instructions (documented, no helper scripts).
  - [ ] Code [Brent]: Add baseline thresholds for indexing and decomposition runtime in a documented matrix.
  - [ ] Code [Brent]: Add fixture metadata (file count, total size, language mix) for repeatable benchmarks.
  - [ ] Docs [Brent]: Add scale test runbook and environment notes.
  - [ ] Tests (Behave) [Brent]: Add scale test scenarios validating thresholds.
  - [ ] Tests (Robot) [Brent]: Add large-project Robot tests for performance runs.
  - [ ] Tests (ASV) [Brent]: Add `asv/benchmarks/scale_fixture_bench.py` for baseline performance.
  - [ ] Quality [Brent]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Brent]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Brent]: `git add .` (only after coverage check passes)
  - [ ] Commit [Brent]: `git commit -m "test(perf): add scale test fixtures"`.
  - [ ] Forgejo PR [Brent]: Open PR from `feature/m6-perf-scale` to `master` with description "Add scale test fixtures and performance baselines.".
  - [ ] Git [Brent]: `git checkout master`
  - [ ] Git [Brent]: `git branch -d feature/m6-perf-scale`

---

### Section 11: Security & Safety [WORKSTREAM F - Luis + Brent]

**Target: Throughout project, critical items by Day 14**

**Note**: Security tasks focus on runtime protections; quality gates are handled in Section 0.

**Parallel Group SEC1: Remove eval() usage [Luis]**
- [ ] **COMMIT (Owner: Luis | Group: SEC1.eval | Branch: feature/m4-security-eval) - Commit message: "fix(security): remove eval-based config parsing"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m4-security-eval`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [X] Code [Luis]: Audit and remove all `eval`/`exec`/`compile` usage from production config paths.
  - [ ] Code [Luis]: Replace any dynamic expression parsing with YAML/JSON parsing and explicit schema validation.
  - [ ] Code [Luis]: Add a hard error if config files contain inline Python or templating directives.
  - [ ] Code [Luis]: Add config scanner that flags disallowed tokens and reports file+line in errors.
  - [ ] Docs [Luis]: Add `docs/reference/security_eval.md` with replacement patterns.
  - [X] Tests (Behave) [Luis]: Add `features/security_eval.feature` scenarios. (completed by Luis)
  - [ ] Tests (Robot) [Luis]: Add `robot/security_eval.robot` smoke tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/security_eval_bench.py` for config parsing baseline.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%. (Code review by Brent still pending)
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "fix(security): remove eval-based config parsing"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m4-security-eval` to `master` with description "Remove eval-based config parsing and add security checks.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m4-security-eval`

**Parallel Group SEC2: Template Injection Prevention [Luis]**
- [ ] **COMMIT (Owner: Luis | Group: SEC2.template | Branch: feature/m4-security-template) - Commit message: "fix(security): harden template rendering"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m4-security-template`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Replace unsafe template usage with a sandboxed renderer and strict token set.
  - [ ] Code [Luis]: Deny attribute access, function calls, and filters; allow only `{var}` substitution with a fixed allowlist.
  - [ ] Code [Luis]: Add max template length + max render output size checks with explicit errors.
  - [ ] Code [Luis]: Add template allowlist validation for each template context key and log rejected keys.
  - [ ] Code [Luis]: Add unit helper to pre-validate template strings at action/plan creation time.
  - [ ] Docs [Luis]: Add `docs/reference/template_security.md` with safe patterns.
  - [ ] Tests (Behave) [Luis]: Add `features/security_templates.feature` scenarios.
  - [ ] Tests (Robot) [Luis]: Add `robot/security_templates.robot` smoke tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/security_template_bench.py` for render baseline.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "fix(security): harden template rendering"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m4-security-template` to `master` with description "Harden template rendering and add tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m4-security-template`

**Parallel Group SEC3: Exception Handling Audit [Luis]**
- [ ] **COMMIT (Owner: Luis | Group: SEC3.exceptions | Branch: feature/m4-security-exceptions) - Commit message: "fix(security): enforce explicit exception handling"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m4-security-exceptions`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Replace silent exception handling with explicit errors and context propagation.
  - [ ] Code [Luis]: Add structured error types for config, provider, and file I/O failures and include plan_id/project_name in error details.
  - [ ] Code [Luis]: Ensure unexpected exceptions are wrapped in `CleverAgentsError` with a safe, user-facing message.
  - [ ] Code [Luis]: Add error code mapping table (error_code -> HTTP-like category) for CLI output consistency.
  - [ ] Code [Luis]: Ensure error details are redacted for secrets before logging.
  - [ ] Docs [Luis]: Document error propagation standards and logging rules.
  - [ ] Tests (Behave) [Luis]: Add `features/security_exceptions.feature` scenarios.
  - [ ] Tests (Robot) [Luis]: Add exception handling integration smoke tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/security_exception_bench.py` for error path overhead baseline.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "fix(security): enforce explicit exception handling"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m4-security-exceptions` to `master` with description "Enforce explicit exception handling and error mapping with tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m4-security-exceptions`

**Parallel Group SEC4: Async Lifecycle Correctness [Luis]**
- [ ] **COMMIT (Owner: Luis | Group: SEC4.async | Branch: feature/m4-security-async-cleanup) - Commit message: "fix(security): close async resources and leaks"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m4-security-async-cleanup`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Close async resources, checkpoint files, and subscription leaks with retention policies.
  - [ ] Code [Luis]: Add graceful cancellation handling to ensure in-flight tasks are awaited and cleaned up.
  - [ ] Code [Luis]: Add a finalizer hook in async services that logs any leaked resources by name.
  - [ ] Code [Luis]: Add cleanup for pending async jobs on shutdown (mark cancelled and persist reason).
  - [ ] Code [Luis]: Add time-bounded shutdown sequence with explicit warnings for forced termination.
  - [ ] Docs [Luis]: Add `docs/reference/async_safety.md` on cleanup rules.
  - [ ] Tests (Behave) [Luis]: Add `features/security_async.feature` scenarios.
  - [ ] Tests (Robot) [Luis]: Add async cleanup integration tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/security_async_cleanup_bench.py` for cleanup overhead baseline.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "fix(security): close async resources and leaks"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m4-security-async-cleanup` to `master` with description "Add async lifecycle cleanup and leak prevention with tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m4-security-async-cleanup`

**Parallel Group SEC5: Secrets Management [Hamza]**
- [ ] **COMMIT (Owner: Hamza | Group: SEC5.secrets | Branch: feature/m4-security-secrets) - Commit message: "feat(security): add secrets masking and validation"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m4-security-secrets`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Mask credentials in logs, validate required keys, and block secret leakage in outputs.
  - [ ] Code [Hamza]: Add a centralized redaction utility (token patterns + config keys) and integrate into CLI output formatting.
  - [ ] Code [Hamza]: Add `--show-secrets` guard flag (default off) for diagnostics that would otherwise print masked fields.
  - [ ] Code [Hamza]: Add secret detection for common key patterns (API keys, tokens) in config and env.
  - [ ] Code [Hamza]: Add redaction for error.details and validation outputs before printing/logging.
  - [ ] Docs [Hamza]: Add `docs/reference/secrets_handling.md`.
  - [ ] Docs [Hamza]: Include examples of masked outputs and `--show-secrets` usage.
  - [ ] Tests (Behave) [Hamza]: Add `features/security_secrets.feature` scenarios.
  - [ ] Tests (Robot) [Hamza]: Add secrets handling integration smoke tests.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/security_secrets_bench.py` for masking overhead baseline.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(security): add secrets masking and validation"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m4-security-secrets` to `master` with description "Add secrets masking/validation and tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m4-security-secrets`

**Parallel Group SEC6: Read-Only Enforcement [Luis]**
- [ ] **COMMIT (Owner: Luis | Group: SEC6.readonly | Branch: feature/m4-security-readonly) - Commit message: "feat(security): enforce read-only actions"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m4-security-readonly`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Validate read-only actions only use read-only skills at execution time.
  - [ ] Code [Luis]: Block write-capable tools in ToolRuntime when plan/action is read-only and include tool name in error.
  - [ ] Code [Luis]: Add read-only enforcement to SkillContext and ChangeSet builder to prevent write artifacts.
  - [ ] Code [Luis]: Add read-only enforcement in CLI commands that would mutate resources (fail fast before execution).
  - [ ] Code [Luis]: Add tests for read-only enforcement on file and git tool calls.
  - [ ] Docs [Luis]: Add `docs/reference/read_only_actions.md`.
  - [ ] Tests (Behave) [Luis]: Add `features/security_readonly.feature` scenarios.
  - [ ] Tests (Robot) [Luis]: Add read-only enforcement integration tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/security_readonly_bench.py` for enforcement overhead baseline.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(security): enforce read-only actions"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m4-security-readonly` to `master` with description "Enforce read-only actions and tool gating with tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m4-security-readonly`
  - [ ] Note: Safety profile enforcement is deferred; see Section 18 POST.safety.

**Parallel Group SEC7: Audit Logging [Hamza]**
- [ ] **COMMIT (Owner: Hamza | Group: SEC7.audit | Branch: feature/m4-security-audit) - Commit message: "feat(security): add audit logging for apply"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m4-security-audit`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add audit log model, migration, and `agents audit list` CLI command.
  - [ ] Code [Hamza]: Record apply start/end events with plan_id, actor, resource list, and changeset hash.
  - [ ] Code [Hamza]: Add CLI filters for `--plan`, `--project`, and `--since` to limit audit output.
  - [ ] Code [Hamza]: Add `audit show <audit_id>` command to view a single entry with full metadata.
  - [ ] Code [Hamza]: Add retention policy for audit logs (configurable days) and prune job.
  - [ ] Docs [Hamza]: Add `docs/reference/audit_logging.md`.
  - [ ] Docs [Hamza]: Document audit retention settings and CLI filter semantics.
  - [ ] Tests (Behave) [Hamza]: Add `features/security_audit.feature` scenarios.
  - [ ] Tests (Robot) [Hamza]: Add audit logging integration tests.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/security_audit_bench.py` for log write overhead baseline.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(security): add audit logging for apply"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m4-security-audit` to `master` with description "Add audit logging for apply with tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m4-security-audit`

### Section 12: Provider Fixes & Runtime Tweaks [WORKSTREAM G - Hamza]

**Target: Days 8-12**

**Parallel Group PROV1: Provider Fixes [Luis]**
- [ ] **COMMIT (Owner: Luis | Group: PROV1.fixes | Branch: feature/m4-provider-fixes) - Commit message: "fix(provider): remove FakeListLLM defaults"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m4-provider-fixes`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Remove FakeListLLM fallback, fix auto-debug provider usage, and implement provider auto-detection.
  - [ ] Code [Luis]: Update settings validation to fail fast when no providers are configured and no mock flag is set.
  - [ ] Code [Luis]: Add explicit `core.mock_providers` flag and block accidental use in non-test mode.
  - [ ] Code [Luis]: Add provider selection trace logging (chosen provider + reason) for diagnostics output.
  - [ ] Code [Luis]: Remove any default provider auto-wiring in container that bypasses settings validation.
  - [ ] Code [Luis]: Add unit helper to resolve provider by name and emit explicit error when not configured.
  - [ ] Docs [Luis]: Update provider configuration docs and error messages.
  - [ ] Docs [Luis]: Add migration note for removing FakeListLLM fallback and mock flag usage.
  - [ ] Tests (Behave) [Luis]: Add `features/provider_fixes.feature` scenarios.
  - [ ] Tests (Robot) [Luis]: Add provider detection smoke tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/provider_selection_bench.py` for provider resolution baseline.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "fix(provider): remove FakeListLLM defaults"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m4-provider-fixes` to `master` with description "Fix provider defaults and enforce configuration validation.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m4-provider-fixes`

**Parallel Group PROV2: Cost Controls & Fallback [Luis]**
- [ ] **COMMIT (Owner: Luis | Group: PROV2.costs | Branch: feature/m4-provider-costs) - Commit message: "feat(provider): add cost controls and fallback"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m4-provider-costs`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Track tokens/costs, enforce budgets, rate limits, and provider fallback order.
  - [ ] Code [Luis]: Add cost tracking fields to plan execution metadata and surface in `plan status`.
  - [ ] Code [Luis]: Add config keys for `budget_per_plan`, `budget_per_day`, and `fallback_providers` with validation.
  - [ ] Code [Luis]: Emit warnings when budget is within 10% of limit and block when exceeded.
  - [ ] Code [Luis]: Add per-provider cost table and default token cost estimates for offline reporting.
  - [ ] Code [Luis]: Add fallback selection logic that skips providers without required capabilities (tool calling, streaming).
  - [ ] Code [Luis]: Persist budget exhaustion events in plan metadata for auditability.
  - [ ] Docs [Luis]: Add `docs/reference/cost_controls.md` with config keys and thresholds.
  - [ ] Tests (Behave) [Luis]: Add `features/cost_controls.feature` scenarios.
  - [ ] Tests (Robot) [Luis]: Add cost control integration smoke tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/cost_controls_bench.py` for cost check overhead.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(provider): add cost controls and fallback"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m4-provider-costs` to `master` with description "Add cost controls, budgets, and provider fallback with tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m4-provider-costs`

---

### Section 13: Additional CLI Commands & UX [Days 10-14]

**Parallel Group CLI0: Core System Commands [Hamza]**
- [ ] **COMMIT (Owner: Hamza | Group: CLI0.core | Branch: feature/m4-cli-core) - Commit message: "feat(cli): add version/info/diagnostics"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m4-cli-core`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Implement `version`, `info`, and `diagnostics` commands with rich/plain/json/yaml output parity.
  - [ ] Code [Hamza]: Add diagnostics checks for config file, database, providers, and filesystem permissions per spec.
  - [ ] Code [Hamza]: Include build metadata in `version` (semver, git sha, build date) and expose in JSON/YAML outputs.
  - [ ] Code [Hamza]: Add `diagnostics --check` to exit non-zero when any check fails; report total pass/fail counts.
  - [ ] Code [Hamza]: Add `diagnostics --format json/yaml` output schema with explicit check names and statuses.
  - [ ] Code [Hamza]: Add detection for missing data dir and invalid config path with actionable error messages.
  - [ ] Docs [Hamza]: Update CLI reference with core system commands and sample outputs.
  - [ ] Docs [Hamza]: Add diagnostics check list and expected remediation steps.
  - [ ] Tests (Behave) [Hamza]: Add `features/cli_core.feature` scenarios for each command output.
  - [ ] Tests (Robot) [Hamza]: Add core command smoke tests.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/cli_core_bench.py` for command runtime baseline.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(cli): add version/info/diagnostics"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m4-cli-core` to `master` with description "Add core CLI system commands and diagnostics with tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m4-cli-core`

---

### Section 14: Concurrency & Cleanup [Days 12-14]

**Parallel Group CONC1: Plan Locking [Luis]**
- [ ] **COMMIT (Owner: Luis | Group: CONC1.lock | Branch: feature/m4-concurrency-locks) - Commit message: "feat(concurrency): add plan and project locks"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m4-concurrency-locks`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Implement plan-level and project-level locks with timeouts.
  - [ ] Code [Luis]: Add `locks` table with owner_id, resource_type, resource_id, acquired_at, expires_at.
  - [ ] Code [Luis]: Ensure locks are enforced in PlanLifecycleService transitions and SubplanService scheduling.
  - [ ] Code [Luis]: Add lock renewal for long-running phases and release locks on graceful shutdown.
  - [ ] Code [Luis]: Allow re-entrant lock acquisition for the same owner and reject conflicting owners with explicit error.
  - [ ] Code [Luis]: Add lock cleanup routine to purge expired locks on startup.
  - [ ] Code [Luis]: Add `agents diagnostics` check to report stale locks count.
  - [ ] Docs [Luis]: Add `docs/reference/concurrency.md` with lock behavior.
  - [ ] Docs [Luis]: Document lock TTL defaults and renewal strategy.
  - [ ] Tests (Behave) [Luis]: Add `features/concurrency.feature` scenarios for lock contention and expiry.
  - [ ] Tests (Robot) [Luis]: Add lock integration smoke tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/concurrency_lock_bench.py` for lock overhead baseline.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(concurrency): add plan and project locks"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m4-concurrency-locks` to `master` with description "Add plan/project locking and diagnostics with tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m4-concurrency-locks`

**Parallel Group CONC2: Resumable Execution [Luis]**
- [ ] **COMMIT (Owner: Luis | Group: CONC2.resume | Branch: feature/m4-concurrency-resume) - Commit message: "feat(concurrency): add plan resume"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m4-concurrency-resume`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Persist step-level progress and implement `plan resume` with graceful shutdown handling.
  - [ ] Code [Luis]: Add resume checkpoints tied to decision IDs and sandbox checkpoints.
  - [ ] Code [Luis]: Validate resume eligibility (non-terminal plans only) and emit clear error for invalid states.
  - [ ] Code [Luis]: Add `plan resume --dry-run` to show where execution will resume without changing state.
  - [ ] Code [Luis]: Add resume metadata in plan (last_completed_step, last_checkpoint_id).
  - [ ] Code [Luis]: Add CLI output for resume summary (phase, step, decision_id) before executing.
  - [ ] Docs [Luis]: Update plan lifecycle docs for resume behavior.
  - [ ] Docs [Luis]: Add resume flow example and error cases.
  - [ ] Tests (Behave) [Luis]: Add `features/plan_resume.feature` scenarios.
  - [ ] Tests (Robot) [Luis]: Add resume integration tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/plan_resume_bench.py` for resume overhead baseline.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(concurrency): add plan resume"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m4-concurrency-resume` to `master` with description "Add plan resume workflow and tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m4-concurrency-resume`

**Parallel Group CONC3: Garbage Collection [Hamza]**
- [ ] **COMMIT (Owner: Hamza | Group: CONC3.gc | Branch: feature/m4-concurrency-cleanup) - Commit message: "feat(ops): add cleanup commands"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m4-concurrency-cleanup`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add cleanup for sandboxes, checkpoints, and stale sessions with CLI commands.
  - [ ] Code [Hamza]: Add retention policy settings for sandbox age, checkpoint count, and session inactivity.
  - [ ] Code [Hamza]: Add `cleanup --dry-run` output (counts + paths) and `--all` override for full purge.
  - [ ] Code [Hamza]: Ensure cleanup skips active sandboxes and running plans (log skipped items).
  - [ ] Code [Hamza]: Add per-resource cleanup summaries (sandboxes/checkpoints/sessions) to output.
  - [ ] Code [Hamza]: Add config keys for cleanup scheduling (manual-only in MVP; stub schedule flag).
  - [ ] Docs [Hamza]: Document cleanup commands and retention defaults.
  - [ ] Docs [Hamza]: Include example `cleanup --dry-run` output and warnings.
  - [ ] Tests (Behave) [Hamza]: Add `features/garbage_collection.feature` scenarios.
  - [ ] Tests (Robot) [Hamza]: Add cleanup integration smoke tests.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/cleanup_bench.py` for cleanup overhead baseline.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(ops): add cleanup commands"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m4-concurrency-cleanup` to `master` with description "Add cleanup commands and retention policies with tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m4-concurrency-cleanup`

---

### Section 15: ACMS & UKO (Advanced Context Management) [Post-M6]

**Target: Begin scaffolding by Day 25; full implementation after M6 (local mode only)**

**Parallel Group ACMS1: UKO Ontology Scaffold [Hamza]**
- [ ] **COMMIT (Owner: Hamza | Group: ACMS1.uko | Branch: feature/m6-acms-uko-schema) - Commit message: "feat(uko): add UKO ontology scaffolding"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m6-acms-uko-schema`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add UKO Layer 0-3 ontology skeleton (RDF/TTL) under `docs/ontology/uko.ttl` with version headers.
  - [ ] Code [Hamza]: Define base URI, version IRI, and prefix conventions for layers and language-specific nodes.
  - [ ] Code [Hamza]: Add minimal Layer 0 nodes (Resource, Artifact, CodeArtifact, Document) to allow end-to-end parsing.
  - [ ] Code [Hamza]: Add Python loaders for UKO nodes (URI parsing, inheritance, versioning metadata).
  - [ ] Code [Hamza]: Add a validation pass that rejects undefined prefixes and missing `rdf:type` for nodes.
  - [ ] Code [Hamza]: Add ontology version registry (current version + supported versions) for future migrations.
  - [ ] Code [Hamza]: Add unit helper to resolve layer inheritance (Layer 3 -> 2 -> 1 -> 0).
  - [ ] Docs [Hamza]: Add `docs/reference/uko.md` describing layers, URI format, and extension points.
  - [ ] Docs [Hamza]: Add a minimal example ontology snippet and parsing walkthrough.
  - [ ] Tests (Behave) [Hamza]: Add `features/uko_ontology.feature` for URI parsing and inheritance checks.
  - [ ] Tests (Robot) [Hamza]: Add `robot/uko_ontology.robot` smoke tests for ontology load.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/uko_load_bench.py` for load throughput.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(uko): add UKO ontology scaffolding"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m6-acms-uko-schema` to `master` with description "Add UKO ontology scaffolding and loaders with tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m6-acms-uko-schema`

**Parallel Group ACMS2: Context Request Protocol [Jeff]** (depends on ACMS1 + CTX1)
- [ ] **COMMIT (Owner: Jeff | Group: ACMS2.crp | Branch: feature/m6-acms-crp-models) - Commit message: "feat(acms): add context request protocol models"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m6-acms-crp-models`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Add `ContextRequest`, `ContextFragment`, `DetailLevel`, and `ContextBudget` models with validation.
  - [ ] Code [Jeff]: Add built-in `builtin/context` skill with tools `request_context`, `query_history`, and `get_context_budget` (stubbed to ACMS).
  - [ ] Code [Jeff]: Add `ContextRequest` fields for `focus`, `breadth`, `depth`, `strategy`, `temporal_scope`, and `skeleton_ratio` with defaulting rules.
  - [ ] Code [Jeff]: Add `ContextFragment` metadata (uko_uri, provenance, relevance_score, token_count, detail_level) and enforce token_count >= 0.
  - [ ] Code [Jeff]: Add `DetailLevelMap` registry with name->integer resolution and inheritance support.
  - [ ] Code [Jeff]: Add `ContextRequest` validation for invalid depth names and unknown strategies.
  - [ ] Docs [Jeff]: Add `docs/reference/crp.md` with request fields, detail levels, and examples.
  - [ ] Docs [Jeff]: Document `DetailLevelMap` resolution rules and defaults.
  - [ ] Tests (Behave) [Jeff]: Add `features/crp_models.feature` for validation and serialization ordering.
  - [ ] Tests (Robot) [Jeff]: Add `robot/crp_models.robot` smoke tests.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/crp_model_bench.py` for validation throughput.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(acms): add context request protocol models"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m6-acms-crp-models` to `master` with description "Add context request protocol models and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m6-acms-crp-models`

**Parallel Group ACMS3: Context Strategies & Registry [Hamza]** (depends on CTX1/CTX2)
- [ ] **COMMIT (Owner: Hamza | Group: ACMS3.strategy | Branch: feature/m6-acms-strategy-registry) - Commit message: "feat(acms): add context strategy registry"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m6-acms-strategy-registry`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Define `ContextStrategy` interface and StrategyRegistry with plugin discovery.
  - [ ] Code [Hamza]: Add stub strategies (keyword, semantic, graph, temporal) with feature flags and no-op defaults.
  - [ ] Code [Hamza]: Add `ContextStrategyResult` model (fragments, stats, errors) with deterministic ordering.
  - [ ] Code [Hamza]: Add configuration to enable/disable strategies per project context policy.
  - [ ] Code [Hamza]: Add per-strategy timeout and max-fragment limits in registry config.
  - [ ] Code [Hamza]: Add registry validation that strategies declare supported resource types.
  - [ ] Docs [Hamza]: Add `docs/reference/context_strategies.md` with strategy contracts and outputs.
  - [ ] Docs [Hamza]: Document strategy config keys (timeouts, max fragments, enable/disable).
  - [ ] Tests (Behave) [Hamza]: Add `features/context_strategy_registry.feature` for registration and selection.
  - [ ] Tests (Robot) [Hamza]: Add `robot/context_strategy_registry.robot` smoke tests.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/context_strategy_bench.py` for registry lookup.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(acms): add context strategy registry"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m6-acms-strategy-registry` to `master` with description "Add context strategy registry and tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m6-acms-strategy-registry`

**Parallel Group ACMS4: Strategy Coordinator & Fusion Engine [Jeff]** (depends on ACMS2/ACMS3)
- [ ] **COMMIT (Owner: Jeff | Group: ACMS4.fusion | Branch: feature/m6-acms-fusion-engine) - Commit message: "feat(acms): add strategy coordinator and fusion engine"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m6-acms-fusion-engine`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Implement StrategyCoordinator with parallel execution and budget allocation.
  - [ ] Code [Jeff]: Implement FusionEngine to dedupe fragments, resolve detail conflicts, and pack within budget.
  - [ ] Code [Jeff]: Allocate budget proportionally by strategy confidence and enforce per-strategy max caps.
  - [ ] Code [Jeff]: Add greedy knapsack packing with deterministic tie-breakers (relevance, detail level, token count).
  - [ ] Code [Jeff]: Add fragment dedupe by UKO URI + hash of rendered text to avoid duplicates across strategies.
  - [ ] Code [Jeff]: Add budget overage guard to drop lowest-relevance fragments and emit warnings.
  - [ ] Docs [Jeff]: Add `docs/reference/acms_fusion.md` with flow diagrams and budget semantics.
  - [ ] Docs [Jeff]: Document StrategyCoordinator execution order and error handling.
  - [ ] Tests (Behave) [Jeff]: Add `features/acms_fusion.feature` for dedupe and budget enforcement.
  - [ ] Tests (Robot) [Jeff]: Add `robot/acms_fusion.robot` integration smoke tests.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/acms_fusion_bench.py` for fusion runtime.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(acms): add strategy coordinator and fusion engine"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m6-acms-fusion-engine` to `master` with description "Add strategy coordinator and fusion engine with tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m6-acms-fusion-engine`

**Parallel Group ACMS5: Scoped Backend Views [Hamza]** (depends on B2.project + CTX1)
- [ ] **COMMIT (Owner: Hamza | Group: ACMS5.scoped | Branch: feature/m6-acms-scoped-view) - Commit message: "feat(acms): add scoped backend view filtering"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m6-acms-scoped-view`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Implement ScopedBackendView that filters text/vector/graph queries to project resources.
  - [ ] Code [Hamza]: Add enforcement hooks in context retrieval services and StrategyCoordinator.
  - [ ] Code [Hamza]: Add allowlist/denylist resolution for resource scopes using project resource links and aliases.
  - [ ] Code [Hamza]: Log and reject context requests that reference resources outside scope (security guard).
  - [ ] Code [Hamza]: Add unit helper to resolve alias -> resource_id mapping and validate uniqueness.
  - [ ] Code [Hamza]: Add guard for mixed project scopes (explicit error when request spans unlinked projects).
  - [ ] Docs [Hamza]: Add `docs/reference/scoped_backend_view.md` with security guarantees.
  - [ ] Docs [Hamza]: Add examples of allowlist/denylist configurations.
  - [ ] Tests (Behave) [Hamza]: Add `features/scoped_view.feature` for cross-project isolation.
  - [ ] Tests (Robot) [Hamza]: Add `robot/scoped_view.robot` smoke tests.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/scoped_view_bench.py` for filter overhead.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(acms): add scoped backend view filtering"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m6-acms-scoped-view` to `master` with description "Add scoped backend view filtering with tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m6-acms-scoped-view`

**Parallel Group ACMS6: Skeleton Compression [Jeff]** (depends on ACMS4 + G4 context tiers)
- [ ] **COMMIT (Owner: Jeff | Group: ACMS6.skeleton | Branch: feature/m6-acms-skeleton-compress) - Commit message: "feat(acms): add skeleton compressor"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m6-acms-skeleton-compress`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Implement skeleton compressor that produces compressed inherited context with `skeleton_ratio`.
  - [ ] Code [Jeff]: Integrate skeleton output into subplan context inheritance and strategy coordinator budgets.
  - [ ] Code [Jeff]: Add stable ordering for compressed fragments to avoid non-deterministic context payloads.
  - [ ] Code [Jeff]: Persist skeleton metadata (ratio, token counts, source decision IDs) for auditability.
  - [ ] Code [Jeff]: Add `skeleton_ratio` validation (0.0-1.0) and default handling per plan.
  - [ ] Code [Jeff]: Add compression summary (original tokens vs compressed tokens) stored in plan metadata.
  - [ ] Docs [Jeff]: Add `docs/reference/skeleton_compressor.md` with ratios and examples.
  - [ ] Docs [Jeff]: Add example of skeleton output for a multi-decision plan.
  - [ ] Tests (Behave) [Jeff]: Add `features/skeleton_compressor.feature` for compression thresholds.
  - [ ] Tests (Robot) [Jeff]: Add `robot/skeleton_compressor.robot` smoke tests.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/skeleton_compressor_bench.py` for compression overhead.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(acms): add skeleton compressor"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m6-acms-skeleton-compress` to `master` with description "Add skeleton compressor and tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m6-acms-skeleton-compress`

---

### Section 16: Context Indexing [Days 15-17]

**Parallel Group CTX1: Repository Indexing [Hamza]**
- [ ] **COMMIT (Owner: Hamza | Group: CTX1.index | Branch: feature/m4-context-indexing) - Commit message: "feat(context): add repo indexing service"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m4-context-indexing`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Implement indexing service with file tree, language detection, and incremental refresh.
  - [ ] Code [Hamza]: Add index metadata table (resource_id, indexed_at, file_count, token_estimate).
  - [ ] Code [Hamza]: Enforce max file size and total size limits from project context policy.
  - [ ] Code [Hamza]: Add file hashing (content hash + mtime) to skip unchanged files during incremental refresh.
  - [ ] Code [Hamza]: Persist per-file index records (path, language, hash, token_count) with indexes on resource_id/path.
  - [ ] Code [Hamza]: Add `IndexingService.refresh(resource_id)` that returns summary stats (added/updated/removed counts).
  - [ ] Code [Hamza]: Add file ignore handling using project include/exclude globs and default ignore list.
  - [ ] Code [Hamza]: Add language detection fallback (by extension when content sniff fails).
  - [ ] Docs [Hamza]: Add `docs/reference/context_indexing.md`.
  - [ ] Docs [Hamza]: Document index storage schema and refresh behavior.
  - [ ] Tests (Behave) [Hamza]: Add `features/context_indexing.feature` scenarios.
  - [ ] Tests (Robot) [Hamza]: Add indexing integration tests.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/context_indexing_bench.py` for indexing throughput baseline.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(context): add repo indexing service"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m4-context-indexing` to `master` with description "Add repository indexing service and tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m4-context-indexing`

**Parallel Group CTX2: Embedding Index [Hamza]**
- [ ] **COMMIT (Owner: Hamza | Group: CTX2.embedding | Branch: feature/m4-context-embedding) - Commit message: "feat(context): add optional embedding search"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m4-context-embedding`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add embedding-based search with opt-in flag and fallback to full-text search.
  - [ ] Code [Hamza]: Add embedding index metadata and cache invalidation on repo updates.
  - [ ] Code [Hamza]: Add embedding provider abstraction (model name, batch size, rate limit) with local-only default.
  - [ ] Code [Hamza]: Store embedding vectors in a local vector store (FAISS or sqlite-vec) and map back to file paths.
  - [ ] Code [Hamza]: Add config keys `context.embedding.enabled`, `context.embedding.model`, and `context.embedding.max_chunks`.
  - [ ] Code [Hamza]: Add chunking strategy (max tokens per chunk, overlap) with deterministic ordering.
  - [ ] Code [Hamza]: Add embedding cache keys and reuse logic to avoid re-embedding unchanged files.
  - [ ] Docs [Hamza]: Add `docs/reference/embedding_search.md`.
  - [ ] Docs [Hamza]: Document embedding storage format and cache invalidation rules.
  - [ ] Tests (Behave) [Hamza]: Add `features/embedding_search.feature` scenarios.
  - [ ] Tests (Robot) [Hamza]: Add embedding search integration tests.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/embedding_search_bench.py` for search runtime baseline.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(context): add optional embedding search"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m4-context-embedding` to `master` with description "Add optional embedding search and tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m4-context-embedding`

---

### Section 17: Skill Registry [Days 17-18]

**Target: Days 17-18 (after Tool/Skill base registry work in C0/C3)**
**Dependencies**: C0.skill.domain + C0.skill.registry + C0.tool.registry + C3.context; MCP refresh depends on C7.mcp.

**Parallel Group SK: Skill Registry Completion [Jeff + Luis + Rui + Aditya]**
  **PARALLEL SUBTRACK SK1.core [Jeff]**: Flattened tool sets, includes, capability summary
  **PARALLEL SUBTRACK SK2.persistence [Luis]**: Persistence extensions for flattened tools
  **PARALLEL SUBTRACK SK3.cli [Aditya]**: CLI `skill tools/refresh` and output parity
  **PARALLEL SUBTRACK SK4.refresh [Aditya]**: MCP refresh hooks and dynamic updates
  **SEQUENTIAL MERGE NOTE**: SK2.persistence must land before SK1.core + SK3.cli to avoid schema mismatches; SK4.refresh depends on C7.mcp and can land after SK1.core.

- [ ] **COMMIT (Owner: Jeff | Group: SK1.core | Branch: feature/m4-skill-registry-core) - Commit message: "feat(skill): add registry flattening and capability summaries"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m4-skill-registry-core`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Implement skill flattening that resolves `includes`, named tool refs, and anonymous inline tools into a deterministic flattened list.
  - [ ] Code [Jeff]: Detect include cycles and emit a clear error listing the cycle path (`skillA -> skillB -> skillA`).
  - [ ] Code [Jeff]: Implement per-tool override merging (shallow merge) with explicit disallow list for non-overridable fields.
  - [ ] Code [Jeff]: Compute capability summary (read/write/checkpointable/idempotent) for each skill from flattened tool metadata.
  - [ ] Code [Jeff]: Add `SkillRegistry.tools(name)` and `SkillRegistry.validate_plan(plan)` to support actor activation and plan validation.
  - [ ] Code [Jeff]: Add deterministic ordering for flattened tools (include order, then tool name).
  - [ ] Code [Jeff]: Add error reporting that includes source file and line when overrides fail.
  - [ ] Docs [Jeff]: Add `docs/reference/skill_registry.md` describing flattening rules, overrides, and validation errors.
  - [ ] Docs [Jeff]: Add example of flattened output with inline tools and includes.
  - [ ] Tests (Behave) [Jeff]: Add `features/skill_registry.feature` scenarios for includes, overrides, cycle detection, and missing tool refs.
  - [ ] Tests (Robot) [Jeff]: Add `robot/skill_registry.robot` for add/list/show/tools flows.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/skill_flatten_bench.py` for flattening throughput.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(skill): add registry flattening and capability summaries"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m4-skill-registry-core` to `master` with description "Add skill registry flattening and capability summaries.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m4-skill-registry-core`

- [ ] **COMMIT (Owner: Luis | Group: SK2.persistence | Branch: feature/m4-skill-registry-db) - Commit message: "feat(skill): persist flattened tool sets"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m4-skill-registry-db`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Add migration fields/tables to store `flattened_tools_json`, `includes_json`, and `capability_summary_json` for each skill.
  - [ ] Code [Luis]: Persist original skill YAML text and computed flattening hash for refresh invalidation.
  - [ ] Code [Luis]: Extend SkillRepository to read/write flattened tool sets and invalidate cached summaries on update.
  - [ ] Code [Luis]: Add uniqueness constraint for skill names and index on namespace for list filters.
  - [ ] Code [Luis]: Add repository method to recompute flattening hash on update and store in DB.
  - [ ] Docs [Luis]: Update `docs/reference/database_schema.md` with skill registry persistence fields.
  - [ ] Docs [Luis]: Add mapping table of persistence fields to domain model properties.
  - [ ] Tests (Behave) [Luis]: Add `features/skill_registry_persistence.feature` scenarios for add/update/refresh persistence.
  - [ ] Tests (Robot) [Luis]: Add repository persistence smoke tests for skills.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/skill_registry_persist_bench.py` for persistence overhead.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(skill): persist flattened tool sets"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m4-skill-registry-db` to `master` with description "Persist flattened skill tool sets and summaries.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m4-skill-registry-db`

- [ ] **COMMIT (Owner: Aditya | Group: SK3.cli | Branch: feature/m4-skill-registry-cli) - Commit message: "feat(cli): add skill tools and refresh commands"**
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git pull origin master`
  - [ ] Git [Aditya]: `git checkout -b feature/m4-skill-registry-cli`
  - [ ] Git [Aditya]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Aditya]: Add `agents skill tools <name>` to show flattened tool list and capability summary.
  - [ ] Code [Aditya]: Add `agents skill refresh <name>|--all` to recompute flattening and sync MCP-backed skills.
  - [ ] Code [Aditya]: Update `skill list/show` output with includes, tool counts, and capability summary fields.
  - [ ] Code [Aditya]: Add `--format json/yaml` output schema for tools/refresh output.
  - [ ] Code [Aditya]: Add CLI errors for refresh when skill not found or MCP sync fails.
  - [ ] Docs [Aditya]: Update CLI reference with skill tools/refresh examples and output columns.
  - [ ] Docs [Aditya]: Document refresh side effects and caching behavior.
  - [ ] Tests (Behave) [Aditya]: Add `features/skill_cli.feature` scenarios for tools/refresh and list/show output.
  - [ ] Tests (Robot) [Aditya]: Add `robot/skill_cli.robot` for CLI smoke tests.
  - [ ] Tests (ASV) [Aditya]: Add `asv/benchmarks/skill_cli_bench.py` for CLI overhead baseline.
  - [ ] Quality [Aditya]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Aditya]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Aditya]: `git add .` (only after coverage check passes)
  - [ ] Commit [Aditya]: `git commit -m "feat(cli): add skill tools and refresh commands"`.
  - [ ] Forgejo PR [Aditya]: Open PR from `feature/m4-skill-registry-cli` to `master` with description "Add skill tools/refresh CLI commands and tests.".
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git branch -d feature/m4-skill-registry-cli`

- [ ] **COMMIT (Owner: Aditya | Group: SK4.refresh | Branch: feature/m4-skill-registry-refresh) - Commit message: "feat(skill): add MCP refresh hooks"**
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git pull origin master`
  - [ ] Git [Aditya]: `git checkout -b feature/m4-skill-registry-refresh`
  - [ ] Git [Aditya]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Aditya]: Implement `SkillRegistry.refresh(name)` and `refresh_all()` to recompute flattened tool sets on demand.
  - [ ] Code [Aditya]: Wire MCP `notifications/tools/list_changed` to trigger skill refresh and registry cache invalidation.
  - [ ] Code [Aditya]: Add safeguard to skip refresh when tool registry is unavailable; emit a single warning with recovery steps.
  - [ ] Code [Aditya]: Add refresh debounce window to avoid repeated refresh storms from MCP notifications.
  - [ ] Code [Aditya]: Add refresh result summary (skills refreshed, failures) for CLI output.
  - [ ] Docs [Aditya]: Add `docs/reference/skill_refresh.md` with MCP refresh behavior and troubleshooting.
  - [ ] Docs [Aditya]: Document debounce defaults and manual refresh usage.
  - [ ] Tests (Behave) [Aditya]: Add `features/skill_refresh.feature` scenarios for MCP update propagation.
  - [ ] Tests (Robot) [Aditya]: Add `robot/skill_refresh.robot` for refresh smoke tests.
  - [ ] Tests (ASV) [Aditya]: Add `asv/benchmarks/skill_refresh_bench.py` for refresh overhead.
  - [ ] Quality [Aditya]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Aditya]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Aditya]: `git add .` (only after coverage check passes)
  - [ ] Commit [Aditya]: `git commit -m "feat(skill): add MCP refresh hooks"`.
  - [ ] Forgejo PR [Aditya]: Open PR from `feature/m4-skill-registry-refresh` to `master` with description "Add MCP refresh hooks and registry updates.".
  - [ ] Git [Aditya]: `git checkout master`
  - [ ] Git [Aditya]: `git branch -d feature/m4-skill-registry-refresh`

---

### Section 18: Deferred Work

Deferred items remain planned but are not part of the 30-day MVP scope.

- [ ] **COMMIT (Owner: Hamza | Group: POST.resource-types.physical | Branch: feature/post-resource-types-physical) - Commit message: "feat(resource): add deferred physical resource types"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/post-resource-types-physical`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add built-in physical resource type YAML configs for git object taxonomy: `git`, `git-remote`, `git-branch`, `git-tag`, `git-commit`, `git-tree`, `git-tree-entry`, `git-stash`, `git-submodule`.
  - [ ] Code [Hamza]: Add filesystem link types: `fs-symlink`, `fs-hardlink` with correct parent/child constraints.
  - [ ] Code [Hamza]: Extend auto-discovery rules for git object graph (bounded depth, filtered by type) and fs link detection.
  - [ ] Docs [Hamza]: Update `docs/reference/resource_types_builtin.md` with deferred physical type flags, parent/child rules, and discovery notes.
  - [ ] Tests (Behave) [Hamza]: Add scenarios ensuring deferred physical types register and validate parent/child constraints.
  - [ ] Tests (Robot) [Hamza]: Add Robot test that lists resource types and asserts the git object taxonomy types.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/resource_type_deferred_physical_bench.py` for registry load overhead.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Git [Hamza]: `git commit -m "feat(resource): add deferred physical resource types"` (only after coverage check passes)
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/post-resource-types-physical` to `master` with description "Add deferred physical resource types (git object taxonomy + fs link types) with discovery rules, docs, and tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/post-resource-types-physical`
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.

- [ ] **COMMIT (Owner: Hamza | Group: POST.resource-types.virtual | Branch: feature/post-resource-types-virtual) - Commit message: "feat(resource): add deferred virtual resource types"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/post-resource-types-virtual`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add built-in virtual resource types: `remote`, `submodule`, `symlink` with equivalence metadata rules.
  - [ ] Code [Hamza]: Update registry bootstrap to include deferred virtual types and hide them from `resource add` scaffolding.
  - [ ] Docs [Hamza]: Update `docs/reference/resource_types_builtin.md` with deferred virtual type descriptions and equivalence notes.
  - [ ] Tests (Behave) [Hamza]: Add scenarios ensuring deferred virtual types exist and remain non-user-addable.
  - [ ] Tests (Robot) [Hamza]: Add Robot test that lists resource types and asserts deferred virtual types.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/resource_type_deferred_virtual_bench.py` for registry load overhead.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Git [Hamza]: `git commit -m "feat(resource): add deferred virtual resource types"` (only after coverage check passes)
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/post-resource-types-virtual` to `master` with description "Add deferred virtual resource types (remote/submodule/symlink) with docs and tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/post-resource-types-virtual`
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.

- [ ] **COMMIT (Owner: Luis | Group: POST.safety-profile | Branch: feature/post-safety-profile) - Commit message: "feat(security): add safety profile model and enforcement stubs"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/post-safety-profile`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Add `SafetyProfile` model with allowed skill categories, sandbox/checkpoint requirements, human-approval flag, and max cost/retry limits per spec.
  - [ ] Code [Luis]: Add `safety_profile` field to Action model and persistence mapping (no legacy compatibility).
  - [ ] Code [Luis]: Add plan-level safety resolution stub (plan > action > project > global) that returns NotImplementedError for enforcement in local mode.
  - [ ] Docs [Luis]: Document safety profile schema, defaults, and stub enforcement behavior in `docs/reference/safety_profile.md`.
  - [ ] Tests (Behave) [Luis]: Add model validation scenarios for safety profile field parsing and constraint validation.
  - [ ] Tests (Robot) [Luis]: Add Robot smoke test that loads a safety profile from YAML and prints serialized output.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/safety_profile_model_bench.py` for validation overhead.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .`
  - [ ] Git [Luis]: `git commit -m "feat(security): add safety profile model and enforcement stubs"`
  - [ ] Forgejo PR [Luis]: Open PR from `feature/post-safety-profile` to `master` with description "Add SafetyProfile model + schema wiring with enforcement stubs for post-30 work.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/post-safety-profile`

- [ ] **COMMIT (Owner: Rui | Group: POST.safety-profile-tests | Branch: feature/post-safety-profile-tests) - Commit message: "test(security): cover safety profile enforcement"**
  - [ ] Git [Rui]: `git checkout master`
  - [ ] Git [Rui]: `git pull origin master`
  - [ ] Git [Rui]: `git checkout -b feature/post-safety-profile-tests`
  - [ ] Git [Rui]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Tests (Behave) [Rui]: Add scenarios for safety profile allow/deny rules and missing profile errors (stub enforcement expected).
  - [ ] Tests (Behave) [Rui]: Add scenarios for cost/retry bounds validation on action creation.
  - [ ] Tests (Robot) [Rui]: Add Robot test that verifies safety profile appears in `action show` output.
  - [ ] Docs [Rui]: Add test fixture notes for safety profile YAML examples in `docs/development/testing.md`.
  - [ ] Tests (ASV) [Rui]: Add `asv/benchmarks/safety_profile_tests_bench.py` for scenario runtime baseline.
  - [ ] Quality [Rui]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Rui]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Rui]: `git add .`
  - [ ] Git [Rui]: `git commit -m "test(security): cover safety profile enforcement"`
  - [ ] Forgejo PR [Rui]: Open PR from `feature/post-safety-profile-tests` to `master` with description "Add Behave/Robot coverage for safety profile validation + stub enforcement.".
  - [ ] Git [Rui]: `git checkout master`
  - [ ] Git [Rui]: `git branch -d feature/post-safety-profile-tests`

- [ ] **COMMIT (Owner: Hamza | Group: POST.resource | Branch: feature/m7-post-resource-equivalence) - Commit message: "feat(resource): add virtual resource equivalence tracking"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m7-post-resource-equivalence`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add `virtual_resource_links` table mapping virtual resource ULID to physical resource ULIDs with uniqueness constraints.
  - [ ] Code [Hamza]: Add `ResourceEquivalenceService` to create/merge virtual resources and update links on content divergence.
  - [ ] Code [Hamza]: Add helper to compute equivalence key (hash or name) for auto-linking during resource discovery.
  - [ ] Code [Hamza]: Add `agents resource equivalence list/add/remove` CLI to manage equivalence sets (post-M6).
  - [ ] Code [Hamza]: Add conflict policy for equivalence (prefer physical over virtual; log divergence events).
  - [ ] Code [Hamza]: Add equivalence reconciliation job to re-hash linked resources on schedule.
  - [ ] Code [Hamza]: Add guard to prevent linking resources of incompatible types (type mismatch error).
  - [ ] Docs [Hamza]: Update `docs/reference/resource_model.md` with physical/virtual equivalence rules and examples.
  - [ ] Docs [Hamza]: Add CLI examples for equivalence list/add/remove.
  - [ ] Tests (Behave) [Hamza]: Add scenarios for linking/unlinking physical resources to virtual resources and divergence updates.
  - [ ] Tests (Robot) [Hamza]: Add Robot test that creates two identical physical resources and verifies a shared virtual resource.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/virtual_resource_bench.py` for equivalence update overhead.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(resource): add virtual resource equivalence tracking"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m7-post-resource-equivalence` to `master` with description "Add virtual resource equivalence tracking and tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m7-post-resource-equivalence`

- [ ] **COMMIT (Owner: Luis | Group: POST.server | Branch: feature/m7-post-server) - Commit message: "feat(client): add server http client"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m7-post-server`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Add HTTP client with health check, version negotiation, and OpenAPI codegen integration.
  - [ ] Code [Luis]: Add config keys for server base URL, API token, and TLS verification; wire into Settings.
  - [ ] Code [Luis]: Map server error responses into domain errors with retry hints.
  - [ ] Code [Luis]: Add per-request timeout + retry policy hooks with exponential backoff for idempotent calls.
  - [ ] Code [Luis]: Add pagination helpers for list endpoints (actions/skills/tools/projects).
  - [ ] Code [Luis]: Add request/response logging with redaction for auth headers.
  - [ ] Code [Luis]: Add TLS verification toggle and explicit warning when disabled.
  - [ ] Docs [Luis]: Add `docs/reference/server_client_http.md` with configuration and connection errors.
  - [ ] Docs [Luis]: Add examples for health check and version negotiation failures.
  - [ ] Tests (Behave) [Luis]: Add scenarios for connection errors and version mismatch handling.
  - [ ] Tests (Robot) [Luis]: Add mock-server connection tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/server_http_client_bench.py` for connection overhead baseline.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(client): add server http client"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m7-post-server` to `master` with description "Add server HTTP client with health checks and tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m7-post-server`

- [ ] **COMMIT (Owner: Luis | Group: POST.server | Branch: feature/m7-post-server) - Commit message: "feat(client): add plan sync and remote execution"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m7-post-server`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Sync actions, request remote plan execution/apply/status, and reconcile remote plan IDs.
  - [ ] Code [Luis]: Add conflict resolution policy (local wins vs server wins) with explicit CLI errors on ambiguity.
  - [ ] Code [Luis]: Add plan sync scope flags (`--actions/--skills/--tools/--projects`) and default to minimal sync.
  - [ ] Code [Luis]: Persist server-side IDs in local metadata for later reconciliation.
  - [ ] Code [Luis]: Add sync summary output (items created/updated/skipped) for CLI display.
  - [ ] Code [Luis]: Add dry-run mode to show what would sync without executing changes.
  - [ ] Docs [Luis]: Document sync semantics and conflict handling in `docs/reference/server_sync.md`.
  - [ ] Docs [Luis]: Add examples for `agents sync --dry-run` outputs.
  - [ ] Tests (Behave) [Luis]: Add scenarios for sync conflicts and retry behavior.
  - [ ] Tests (Robot) [Luis]: Add mock-server sync tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/server_sync_bench.py` for sync throughput baseline.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(client): add plan sync and remote execution"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m7-post-server` to `master` with description "Add plan sync and remote execution with tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m7-post-server`

- [ ] **COMMIT (Owner: Luis | Group: POST.server | Branch: feature/m7-post-server) - Commit message: "feat(client): add websocket updates"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m7-post-server`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Add WebSocket client for plan updates with reconnect/backoff policy.
  - [ ] Code [Luis]: Define event schema mapping for plan status/progress/log stream updates.
  - [ ] Code [Luis]: Add heartbeat/ping handling and resume from last event ID on reconnect.
  - [ ] Code [Luis]: Add event version negotiation and explicit error for incompatible server schema.
  - [ ] Code [Luis]: Add event de-duplication by event_id and ordered delivery guarantees.
  - [ ] Code [Luis]: Add configurable reconnect backoff parameters in settings.
  - [ ] Docs [Luis]: Add `docs/reference/server_websocket.md` with event types and reconnect rules.
  - [ ] Docs [Luis]: Add examples of event payloads and resume behavior.
  - [ ] Tests (Behave) [Luis]: Add scenarios for reconnect and event ordering.
  - [ ] Tests (Robot) [Luis]: Add WebSocket mock tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/server_ws_bench.py` for message handling baseline.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(client): add websocket updates"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m7-post-server` to `master` with description "Add WebSocket updates for server plan events.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m7-post-server`

- [ ] **COMMIT (Owner: Hamza | Group: POST.server | Branch: feature/m7-post-server) - Commit message: "feat(client): add remote project support"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m7-post-server`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add remote resource selection and server execution request wiring.
  - [ ] Code [Hamza]: Add project-name resolution rules for remote namespaces and server aliases.
  - [ ] Code [Hamza]: Add `agents project list --remote` and `plan use --remote` scaffolding (server-only).
  - [ ] Code [Hamza]: Add explicit error when remote project is not found or user lacks access.
  - [ ] Code [Hamza]: Add remote project caching with TTL to minimize repeated server calls.
  - [ ] Code [Hamza]: Add CLI output for remote project list (namespace, id, last_updated).
  - [ ] Docs [Hamza]: Add `docs/reference/server_remote_projects.md` with project selection semantics.
  - [ ] Docs [Hamza]: Add examples for `project list --remote` and `plan use --remote`.
  - [ ] Tests (Behave) [Hamza]: Add scenarios for remote project selection errors.
  - [ ] Tests (Robot) [Hamza]: Add remote execution mock tests.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/server_remote_project_bench.py` for request overhead baseline.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(client): add remote project support"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m7-post-server` to `master` with description "Add remote project support and CLI scaffolding.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m7-post-server`

- [ ] **COMMIT (Owner: Rui | Group: POST.repl | Branch: feature/m7-post-repl) - Commit message: "feat(cli): add interactive repl"**
  - [ ] Git [Rui]: `git checkout master`
  - [ ] Git [Rui]: `git pull origin master`
  - [ ] Git [Rui]: `git checkout -b feature/m7-post-repl`
  - [ ] Git [Rui]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Rui]: Implement `agents repl` command that dispatches to existing CLI commands with shared config handling.
  - [ ] Code [Rui]: Add history support with opt-out (`--no-history`) and default path `~/.cleveragents/history`.
  - [ ] Code [Rui]: Add tab-completion for top-level commands and last command repetition (`!!`).
  - [ ] Code [Rui]: Add prompt context (active project/plan) and graceful handling of Ctrl+C/Ctrl+D.
  - [ ] Code [Rui]: Add multi-line input support for long commands and quoted strings.
  - [ ] Code [Rui]: Add `:help` and `:exit` built-in REPL commands.
  - [ ] Docs [Rui]: Add REPL usage guide with supported commands and exit behavior.
  - [ ] Docs [Rui]: Document history file location and privacy considerations.
  - [ ] Tests (Behave) [Rui]: Add REPL behavior scenarios (history on/off, unknown command, exit).
  - [ ] Tests (Robot) [Rui]: Add REPL smoke tests for command dispatch.
  - [ ] Tests (ASV) [Rui]: Add `asv/benchmarks/repl_bench.py` for REPL startup baseline.
  - [ ] Quality [Rui]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Rui]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Rui]: `git add .` (only after coverage check passes)
  - [ ] Commit [Rui]: `git commit -m "feat(cli): add interactive repl"`.
  - [ ] Forgejo PR [Rui]: Open PR from `feature/m7-post-repl` to `master` with description "Add interactive REPL with history and tests.".
  - [ ] Git [Rui]: `git checkout master`
  - [ ] Git [Rui]: `git branch -d feature/m7-post-repl`

- [ ] **COMMIT (Owner: Luis | Group: POST.auth | Branch: feature/m7-post-auth) - Commit message: "feat(cli): add auth and team commands"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m7-post-auth`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Add `agents auth login/logout/status` and `agents team list/use` commands with stubbed responses when server is disabled.
  - [ ] Code [Luis]: Add config keys for auth token storage, active team, and default namespace (client-only stubs).
  - [ ] Code [Luis]: Wire stubbed commands to `AuthClient` and `ServerClient` interfaces (raise NotImplementedError when no server).
  - [ ] Code [Luis]: Add secure token storage via OS keyring (fallback to encrypted file when unavailable).
  - [ ] Code [Luis]: Add `auth logout` cleanup to remove tokens and clear active team/namespace.
  - [ ] Code [Luis]: Add token format validation and redaction in all outputs.
  - [ ] Docs [Luis]: Document auth/team workflows and local-only stub behavior.
  - [ ] Docs [Luis]: Add examples for login/logout/status and token storage notes.
  - [ ] Tests (Behave) [Luis]: Add auth/team CLI scenarios (stubbed responses, missing server errors).
  - [ ] Tests (Robot) [Luis]: Add auth/team integration smoke tests.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/auth_cli_bench.py` for auth command baseline.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(cli): add auth and team commands"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m7-post-auth` to `master` with description "Add auth and team CLI commands with stubs and tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m7-post-auth`

- [ ] **COMMIT (Owner: Jeff | Group: POST.tui | Branch: feature/m7-post-tui) - Commit message: "feat(ui): add TUI/Web interface"**
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git pull origin master`
  - [ ] Git [Jeff]: `git checkout -b feature/m7-post-tui`
  - [ ] Git [Jeff]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Jeff]: Define UI data-provider interface (plans, sessions, validations, diffs, logs) backed by local services.
  - [ ] Code [Jeff]: Implement minimal TUI with plan list, plan detail, diff viewer, and validation summary panes.
  - [ ] Code [Jeff]: Add Web UI stub that serves the same data via local-only routes (read-only by default).
  - [ ] Code [Jeff]: Add auto-refresh interval config and manual refresh keybinds for TUI.
  - [ ] Docs [Jeff]: Add UI usage guide with navigation and data-refresh behavior.
  - [ ] Tests (Behave) [Jeff]: Add UI behavior scenarios (list, detail, diff, refresh).
  - [ ] Tests (Robot) [Jeff]: Add UI smoke tests for route loading and TUI navigation.
  - [ ] Tests (ASV) [Jeff]: Add `asv/benchmarks/ui_render_bench.py` for UI render baseline.
  - [ ] Quality [Jeff]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Jeff]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Jeff]: `git add .` (only after coverage check passes)
  - [ ] Commit [Jeff]: `git commit -m "feat(ui): add TUI/Web interface"`.
  - [ ] Forgejo PR [Jeff]: Open PR from `feature/m7-post-tui` to `master` with description "Add TUI/Web UI stubs with tests.".
  - [ ] Git [Jeff]: `git checkout master`
  - [ ] Git [Jeff]: `git branch -d feature/m7-post-tui`

- [ ] **COMMIT (Owner: Hamza | Group: POST.dbresources | Branch: feature/m7-post-resource-db) - Commit message: "feat(resource): add database resources"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m7-post-resource-db`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add database resource types (postgres, mysql, sqlite, duckdb) with connection args and auth handling.
  - [ ] Code [Hamza]: Implement sandbox strategy using transaction wrappers and read-only toggles.
  - [ ] Code [Hamza]: Add connection validation and safe error messaging (mask credentials in logs).
  - [ ] Docs [Hamza]: Document database resource configuration and supported auth options.
  - [ ] Tests (Behave) [Hamza]: Add database resource scenarios (connection validation, read-only enforcement).
  - [ ] Tests (Robot) [Hamza]: Add database resource integration tests (local sqlite/duckdb only).
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/db_resource_bench.py` for resource registration baseline.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(resource): add database resources"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m7-post-resource-db` to `master` with description "Add database resource types and tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m7-post-resource-db`

- [ ] **COMMIT (Owner: Hamza | Group: POST.cloud | Branch: feature/m7-post-resource-cloud) - Commit message: "feat(resource): add cloud infrastructure resources"**
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git pull origin master`
  - [ ] Git [Hamza]: `git checkout -b feature/m7-post-resource-cloud`
  - [ ] Git [Hamza]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Hamza]: Add cloud resource types (aws, gcp, azure) with credential fields and region/tenant metadata.
  - [ ] Code [Hamza]: Add stubbed sandbox strategies that validate configuration and return NotImplementedError for execution.
  - [ ] Code [Hamza]: Add credential resolution from environment variables and profile names (no secrets logged).
  - [ ] Docs [Hamza]: Document cloud resource configuration and local-only stub behavior.
  - [ ] Tests (Behave) [Hamza]: Add cloud resource scenarios (schema validation, stub errors).
  - [ ] Tests (Robot) [Hamza]: Add cloud resource integration tests with stubbed responses.
  - [ ] Tests (ASV) [Hamza]: Add `asv/benchmarks/cloud_resource_bench.py` for resource registration baseline.
  - [ ] Quality [Hamza]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Hamza]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Hamza]: `git add .` (only after coverage check passes)
  - [ ] Commit [Hamza]: `git commit -m "feat(resource): add cloud infrastructure resources"`.
  - [ ] Forgejo PR [Hamza]: Open PR from `feature/m7-post-resource-cloud` to `master` with description "Add cloud infrastructure resource stubs and tests.".
  - [ ] Git [Hamza]: `git checkout master`
  - [ ] Git [Hamza]: `git branch -d feature/m7-post-resource-cloud`

- [ ] **COMMIT (Owner: Luis | Group: POST.permissions | Branch: feature/m7-post-permissions) - Commit message: "feat(security): add permission system"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m7-post-permissions`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Implement namespace/project/plan/skill permission model (role bindings, default deny, allow overrides).
  - [ ] Code [Luis]: Add enforcement hooks at CLI/service boundaries (server-only; local mode returns permissive defaults).
  - [ ] Code [Luis]: Add role enums (owner/admin/editor/viewer) and default role mapping for local mode.
  - [ ] Docs [Luis]: Document permission model, role matrix, and server-only behavior.
  - [ ] Tests (Behave) [Luis]: Add permission scenarios (allow/deny, missing role, server disabled).
  - [ ] Tests (Robot) [Luis]: Add permission integration tests with stubbed server client.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/permission_check_bench.py` for enforcement baseline.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(security): add permission system"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m7-post-permissions` to `master` with description "Add permission system stubs and tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m7-post-permissions`

- [ ] **COMMIT (Owner: Luis | Group: POST.safety | Branch: feature/m7-post-safety) - Commit message: "feat(security): add safety profile enforcement"**
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git pull origin master`
  - [ ] Git [Luis]: `git checkout -b feature/m7-post-safety`
  - [ ] Git [Luis]: `git fetch origin && git merge origin/master` (run before final tests and before commit)
  - [ ] Code [Luis]: Add SafetyProfile model, CLI flags, and execution enforcement hooks (server-only for now).
  - [ ] Code [Luis]: Add safety profile resolution order (plan > project > global) with defaults.
  - [ ] Code [Luis]: Add policy validation for forbidden tools/resources and explicit denial messages.
  - [ ] Docs [Luis]: Document safety profile options, defaults, and server-only behavior.
  - [ ] Tests (Behave) [Luis]: Add safety profile enforcement scenarios (deny/allow paths).
  - [ ] Tests (Robot) [Luis]: Add safety profile integration tests with stubbed server client.
  - [ ] Tests (ASV) [Luis]: Add `asv/benchmarks/safety_profile_bench.py` for enforcement baseline.
  - [ ] Quality [Luis]: Run `nox` (all default sessions, including benchmark).
  - [ ] Quality [Luis]: Verify coverage >=97% via `nox -s coverage_report`. If coverage is <97% then review the current unit test coverage report at `build/coverage.xml` and use it to write new Behave based unit tests to improve code coverage. Specifically, write Behave style unit tests that are descriptively named and specifically improves coverage on whichever file has the most uncovered lines by writing tests that will target the uncovered lines in the report. Once that is done rerun `nox -s coverage_report` to verify all tests pass and coverage is above >=97%. Only mark this as complete once coverage is >=97%, if not repeat this task as many times as is needed until coverage reaches >=97%.
  - [ ] Git [Luis]: `git add .` (only after coverage check passes)
  - [ ] Commit [Luis]: `git commit -m "feat(security): add safety profile enforcement"`.
  - [ ] Forgejo PR [Luis]: Open PR from `feature/m7-post-safety` to `master` with description "Add safety profile enforcement stubs and tests.".
  - [ ] Git [Luis]: `git checkout master`
  - [ ] Git [Luis]: `git branch -d feature/m7-post-safety`

---

## Timeline Summary

### TEAM ROLES AND ASSIGNMENTS

| Developer | Role | Primary Focus Areas | Notes |
|-----------|------|---------------------|-------|
| **Jeff** | CTO/Lead Architect | Critical path items, architectural decisions, complex integrations | Fastest, most expert developer - handles blocking issues |
| **Luis** | Senior Python Architect | Domain models, persistence, algorithms, state machines | Good architecture but can be pedantic - needs clear requirements |
| **Aditya** | Domain Expert (Agents/LLMs) | Actor YAML configs, hierarchical actors, skill execution | Understands topic best but code may need cleanup |
| **Hamza** | Python/RDF Expert | Resources, sandbox, database, general Python | Well-rounded, no agent experience - assign infrastructure |
| **Rui** | Fast Developer | Testing (Behave/Robot), simpler implementations (post-2026-02-27) | New to Python; unavailable 2026-02-13 to 2026-02-27 |
| **Brent** | Quality Specialist | Code review, linting, type checking, documentation | Slow but detail-oriented - low contention independent work |
| **Mike/Brian** | Sysadmins | Deployment, infrastructure setup | Minimal coding tasks |

### Week 1 (Days 1-7) - MVP Target (Source Code Only)
| Day | Morning Focus | Owner | Afternoon Focus | Owner |
|-----|---------------|-------|-----------------|-------|
| 1 | A5.alpha + A5.action_arguments DB migrations + ORM models | Jeff + Luis | B1.core Project/Resource models | Hamza |
| 2 | A5.gamma repos/services + A5.tests + B3.cleanup (legacy project CLI) | Jeff + Brent (tests) | B2.persistence/B2.service + B3.cli Resource CLI | Hamza + Brent (tests) |
| 3 | B4.sandbox git_worktree | Jeff + Hamza | B4.sandbox manager + tests | Luis + Brent (tests) |
| 4 | C1.schema/C1.examples Actor YAML | Aditya | C2.loader/C2.compiler + C2.legacy v2 removal | Aditya + Jeff |
| 5 | C3.protocol/C3.context/C3.inline Skill framework | Jeff | C4.file/C4.search Built-in skills | Luis + Jeff |
| 6 | C7.mcp MCP Adapter | Aditya + Jeff | C4.git + C5.model/C5.router Change tracking | Luis + Jeff |
| 7 | C5.diff Diff review artifacts | Luis | C6.pipeline/C6.gating Validation pipeline | Luis + Brent (tests) |

### Week 2 (Days 8-14) - M3 Complete + Plan-Actor Integration
| Day | Focus | Owner | Deliverable |
|-----|-------|-------|-------------|
| 8 | C8.providers + C9.execute Plan-Actor integration | Jeff + Aditya (+Luis support) | Execute phase + providers ready |
| 9 | C9.apply Apply Phase + Review | Jeff + Luis | Apply flow ready with review gates |
| 10 | D1.domain Decision Model | Hamza (+Jeff review) | Decision model committed |
| 11 | D2.service Decision Recording | Hamza (+Jeff review) | Decision recording committed |
| 12 | E1.domain Subplan Model | Luis (+Jeff review) | Subplan domain committed |
| 13 | E2.service/E2.actor Subplan spawning | Jeff + Aditya (+Luis support) | Subplan spawn committed |
| 14 | End-to-end integration testing | All + Brent (tests/QA) | M3 milestone verified |

### Week 3 (Days 15-21) - M4 Target (Decision Tree + Correction)
| Day | Focus | Owner | Deliverable |
|-----|-------|-------|-------------|
| 15 | D5.db/D5.repo Decision persistence | Hamza (+Jeff review) | Decision storage wired |
| 16 | D3.cli Decision viewing | Hamza (+Jeff review) | `agents [--data-dir PATH] [--config-path PATH] plan tree`, `agents [--data-dir PATH] [--config-path PATH] plan explain` |
| 17 | D4.revert Decision correction (revert) | Jeff (+Luis support) | `agents [--data-dir PATH] [--config-path PATH] plan correct` revert flow |
| 18 | D4.append + D5.di Decision wiring | Jeff + Hamza (+Luis support) | Append correction + service wiring |
| 19 | E3.exec Parallel Subplan Execution | Luis (+Jeff review) | Concurrent subplans |
| 20 | E4.merge Result Merging | Jeff (+Luis support) | Git-style merge for subplans |
| 21 | M4 integration testing | All + Rui (tests) + Brent (QA) | Decision correction working |

### Week 4 (Days 22-30) - M6 Target (Large Project Autonomy - LOCAL MODE ONLY)
| Day | Focus | Owner | Deliverable |
|-----|-------|-------|-------------|
| 22-23 | CTX1.index Context indexing + G3.semantic | Hamza + Luis | Large codebase indexing + semantic validation |
| 24-25 | G4.context Hot/Warm/Cold tiers + G2.checkpoint | Hamza + Luis | Three-tier memory + checkpointing |
| 26-27 | G1.decompose + G5.estimate | Jeff + Hamza | Autonomous decomposition + estimation |
| 28 | F0.stubs | Luis | Client stubs (NOT server impl) |
| 29 | M6 perf triage + large project tests | All + Rui (tests) + Brent (QA) | 10K file perf target |
| 30 | M6 integration testing | All + Rui (tests) + Brent (QA) | Large project autonomy verified (Server connectivity DEFERRED) |

> **Note**: Server connectivity (F1-F4) is DEFERRED beyond Day 30. Days 26-29 focus on client **stubs only** and large project testing. The server is a separate project.

### Week 5 (Days 31-35) - M7 Target (Server Connectivity - Client Side Only)
| Day | Focus | Owner | Deliverable |
|-----|-------|-------|-------------|
| 31-32 | F1.client Server client infrastructure | Luis (+Jeff review) | HTTP client for server communication |
| 33 | F2.sync Plan sync client | Luis (+Jeff review) | Client can sync plans to server |
| 34 | F3.ws WebSocket client | Luis (+Jeff review) | Client receives real-time updates |
| 35 | F4.remote Remote project support | Hamza (+Jeff review) | Client can request server execution |

### Week 6 (Days 36-40) - M8 Target (Full Feature Set + Polish)
| Day | Focus | Owner | Deliverable |
|-----|-------|-------|-------------|
| 36 | A6.* automation level refinements | Jeff + Luis | Automation modes stabilized |
| 37 | G5.estimate cost/risk estimation | Hamza (+Jeff review) | Estimation working |
| 38 | G2.checkpoint rollback | Luis (+Jeff review) | Checkpointing + rollback |
| 39 | G1.decompose + G3.semantic performance tuning | Jeff + Luis | Benchmarks passing |
| 40 | Final integration + documentation | All + Rui (tests) + Brent (QA) | Release candidate ready |

### Continuous Tasks (Throughout)

**Brent (Quality - Independent, Low Contention)**:
- Review all PRs within 4 hours of submission
- Run `nox -s typecheck` on all branches before merge
- Run `nox -s lint` and ensure 0 warnings
- Monitor test coverage (must stay >=97%)
- Update documentation for API changes
- Security audit: no eval(), no template injection, no secrets in code

**Rui (Testing - Parallel with Feature Work)**:
- Unavailable 2026-02-13 to 2026-02-27; Brent owns M1/M2 testing during this window
- Write Behave scenarios for each feature (before implementation starts) after return
- Write Robot integration tests for each milestone after return
- Run full test suite daily once back on rotation
- Report test failures immediately
- Maintain test fixtures and mocks in `features/`

### Critical Path Dependencies

```
Day 1: A5 (Persistence + Action Args) ───────────────────────────────┐
Day 2: B1.core/B2.persistence/B2.service/B3.cli (Project/Resource) ─┐│
Day 3: B4.sandbox (Sandbox) ────────────────────────────────────────┼┤
Day 4: C1.schema/C2.legacy/C2.compiler (Actor) ─────────────────────┘│
Day 5: C3.protocol/C4.file (Skills) ─────────────────────────────────┤
Day 6: C4.search/C5.model (Change Tracking) ─────────────────────────┤
Day 7: C6.pipeline/C7.mcp/C8.providers (Validation + Providers) ─────┘
                                                        │
Day 8-9: C9.execute/C9.apply (Integration + Apply) ────────────────┤
Day 10-14: D1-E2 (Decisions + Subplans) ────────────────┤
                                                        │
                          MERGE POINT M3 ◄──────────────┘
                                                        │
Day 15-21: D3-E4 (Correction + Parallel) ───────────────┤
                                                        │
                          MERGE POINT M4 ◄──────────────┘
                                                        │
Day 22-30: CTX/G + F0 (Context + Large Project + Stubs) ─┘
```

### Risk Mitigation

| Risk | Mitigation | Owner |
|------|------------|-------|
| Git worktree complexity | Use B4.sandbox git_worktree with pre-commit verification and isolation tests | Jeff |
| Multi-file generation reliability | Validate C9.execute/C9.apply flows with diff review artifacts and Robot E2E | Luis + Brent (Rui after 2026-02-27) |
| Decision tree correction bugs | Jeff reviews D4 correction + checkpointing; add revert/append Behave coverage | Jeff |
| Large codebase performance | Profile CTX1/CTX2 indexing + hot/warm/cold tiers; enforce bounded memory tests | Luis + Hamza |
| Server connectivity stubs stability | Keep client-only stubs isolated; gate with feature flags and contract tests | Jeff + Luis |

### Definition of Done (Each Task)

1. ✅ Code implemented with full type annotations
2. ✅ Behave scenarios written and passing
3. ✅ Robot integration tests (for CLI/E2E tasks) passing
4. ✅ `nox -s typecheck` passes with 0 errors
5. ✅ `nox -s lint` passes with 0 warnings
6. ✅ Test coverage >=97%
7. ✅ PR reviewed by Brent (or Jeff for critical items)
8. ✅ Implementation notes added to this document

---

## MILESTONE SUCCESS CRITERIA (DETAILED)

### M1: MVP (Day 7) - Minimally Usable for Source Code

**End-to-end verification command sequence:**
```bash
# 1. Create an action
cat > /tmp/test_action.yaml <<EOF
name: local/test-action
description: MVP test action
strategy_actor: openai/gpt-4
execution_actor: openai/gpt-4
definition_of_done: "All files created successfully"
EOF
agents [--data-dir PATH] [--config-path PATH] action create --config /tmp/test_action.yaml
agents [--data-dir PATH] [--config-path PATH] action available local/test-action

# 2. Register a git resource (built-in git-checkout type)
agents [--data-dir PATH] [--config-path PATH] resource add git-checkout local/main-repo \
  --path /path/to/repo \
  --branch main  # Optional; default is main per spec

# 3. Create project and link resource
agents [--data-dir PATH] [--config-path PATH] project create local/test-project
agents [--data-dir PATH] [--config-path PATH] project link-resource local/test-project local/main-repo

# Note: project link-resource is defined in Section 4 (B3.cli)

# 4. Use action to create plan
agents [--data-dir PATH] [--config-path PATH] plan use local/test-action local/test-project

# 5. Execute plan (sandbox created, changes made)
agents [--data-dir PATH] [--config-path PATH] plan execute <plan_id>

# 6. Review diff
agents [--data-dir PATH] [--config-path PATH] plan diff <plan_id>

# 7. Apply changes
agents [--data-dir PATH] [--config-path PATH] plan apply <plan_id>

# 8. Verify changes in repo
cd /path/to/repo && git log -1  # Shows CleverAgents commit
```

**Technical Criteria:**
- Plan and Action records persist to SQLite database.
- Phase transitions (ACTION → STRATEGIZE → EXECUTE → APPLY → APPLIED) work correctly.
- Git worktree sandbox creates isolated working directory.
- Changes in sandbox do not affect original until Apply.
- At least 3 automation levels work (manual mode minimum).
- Error handling produces actionable messages.
- Test coverage remains >=97%.

### M3: Full Plan Lifecycle with Actors (Day 14)

**End-to-end verification:**
```bash
# Create actor YAML file
cat > my_actor.yaml <<EOF
version: "3"
name: my_coder
namespace: local
type: llm
model: openai/gpt-4
temperature: 0.7
system_prompt: "You are a helpful coding assistant."
EOF

# Load actor
agents [--data-dir PATH] [--config-path PATH] actor add --file ./my_actor.yaml

# Create action using custom actor
cat > /tmp/custom_action.yaml <<EOF
name: local/custom-action
strategy_actor: local/my_coder
execution_actor: local/my_coder
definition_of_done: "Code written and tests pass"
EOF
agents [--data-dir PATH] [--config-path PATH] action create --config /tmp/custom_action.yaml --available

# Use and execute (should use custom actor)
agents [--data-dir PATH] [--config-path PATH] plan use local/custom-action local/test-project
agents [--data-dir PATH] [--config-path PATH] plan execute <plan_id>

# Verify multi-file generation
agents [--data-dir PATH] [--config-path PATH] plan artifacts <plan_id>  # Should show multiple files
```

**Technical Criteria:**
- Actor YAML files parse and validate correctly.
- Actors compile to LangGraph StateGraphs.
- Inline skill code executes in sandboxed environment.
- Built-in file skills (read/write/edit/delete) work.
- ChangeSet built from skill invocations (not parsed from output).
- Validation pipeline runs (syntax check, lint, tests).
- Multi-file generation produces correct ChangeSet.
- MCP skill adapter can connect to external servers (basic).

### M4: Decision Tree & Correction (Day 21)

**End-to-end verification:**
```bash
# Execute a plan to generate decisions
agents [--data-dir PATH] [--config-path PATH] plan use local/complex-action local/large-project
agents [--data-dir PATH] [--config-path PATH] plan execute <plan_id>

# View decision tree
agents [--data-dir PATH] [--config-path PATH] plan tree <plan_id>
# Output shows:
# ├── [prompt_definition] "Build user authentication"
# │   ├── [strategy_choice] "Use JWT tokens" (confidence: 0.85)
# │   │   └── [subplan_spawn] "Create token service" → subplan_01ABC
# │   └── [implementation_choice] "Store tokens in Redis"

# Explain a decision
agents [--data-dir PATH] [--config-path PATH] plan explain <decision_id>
# Shows: question, chosen_option, alternatives, rationale, context

# Add and list project invariants
agents [--data-dir PATH] [--config-path PATH] invariant add --project local/large-project "Use session cookies"
agents [--data-dir PATH] [--config-path PATH] invariant list --project local/large-project

# Correct a decision (dry run first)
agents [--data-dir PATH] [--config-path PATH] plan correct <decision_id> --mode=revert --guidance "Use session cookies instead of JWT" --dry-run
# Shows impact: 3 decisions, 1 subplan affected

# Execute correction
agents [--data-dir PATH] [--config-path PATH] plan correct <decision_id> --mode=revert --guidance "Use session cookies instead of JWT"
# Re-executes from that point

# Verify new outcome
agents [--data-dir PATH] [--config-path PATH] plan tree <plan_id>
# Shows corrected decision and regenerated downstream work
```

**Technical Criteria:**
- Decisions recorded during Strategize with full context snapshot.
- Decision tree persists to database.
- `agents [--data-dir PATH] [--config-path PATH] plan tree` displays ASCII tree correctly.
- `agents [--data-dir PATH] [--config-path PATH] plan explain` shows all decision details.
- Correction in revert mode:
  - Archives old decisions.
  - Rolls back sandbox to checkpoint.
  - Re-executes from decision point.
  - Generates new downstream decisions.
- Correction in append mode creates fix subplan.
- History preserved for comparison.

### M5: Subplans & Parallel Execution (Day 25)

**End-to-end verification:**
```bash
# Execute plan that spawns multiple subplans
agents [--data-dir PATH] [--config-path PATH] plan use local/refactor-action local/monorepo
agents [--data-dir PATH] [--config-path PATH] plan execute <plan_id>

# View subplan tree
agents [--data-dir PATH] [--config-path PATH] plan tree <plan_id>
# Shows:
# ├── [root] Refactor codebase
# │   ├── [subplan] Refactor auth module (status: COMPLETE)
# │   ├── [subplan] Refactor api module (status: PROCESSING)
# │   └── [subplan] Refactor utils module (status: QUEUED)

# Check status until complete
agents [--data-dir PATH] [--config-path PATH] plan status <plan_id>

# Verify merged results
agents [--data-dir PATH] [--config-path PATH] plan diff <plan_id>  # Shows merged changes from all subplans
```

**Technical Criteria:**
- SUBPLAN_SPAWN decisions created during Strategize.
- Subplans actually spawned during Execute.
- Sequential subplan execution works (one at a time).
- Parallel subplan execution works (with max_parallel limit).
- Each subplan has isolated sandbox.
- Three-way merge combines non-conflicting changes.
- Merge conflicts detected and marked.
- Parent plan tracks all subplan statuses.
- A plan with 10+ subplans completes successfully.

### M6: Large Project Handling (Day 30)

**End-to-end verification:**
```bash
# Index a large project (10,000+ files)
agents [--data-dir PATH] [--config-path PATH] project create local/large-project
agents [--data-dir PATH] [--config-path PATH] resource add git-checkout local/large-repo \
  --path /path/to/large/repo \
  --branch main  # Optional; default is main per spec
agents [--data-dir PATH] [--config-path PATH] project link-resource local/large-project local/large-repo

# Note: project link-resource is defined in Section 4 (B3.cli)

# Verify indexing handles scale
agents [--data-dir PATH] [--config-path PATH] project show local/large-project
# Shows: 15,247 files indexed, 2.3M tokens

# Execute complex hierarchical task
cat > /tmp/port_action.yaml <<EOF
name: local/port-to-typescript
strategy_actor: local/architect
execution_actor: local/coder
definition_of_done: "All Python files converted to TypeScript with tests"
EOF
agents [--data-dir PATH] [--config-path PATH] action create --config /tmp/port_action.yaml --available

agents [--data-dir PATH] [--config-path PATH] plan use local/port-to-typescript local/large-project
agents [--data-dir PATH] [--config-path PATH] plan execute <plan_id>

# Monitor hierarchical decomposition
agents [--data-dir PATH] [--config-path PATH] plan tree <plan_id>
# Shows multi-level hierarchy:
# ├── Root: Port codebase to TypeScript
# │   ├── Phase 1: Core modules (3 subplans)
# │   │   ├── Port auth module (5 subplans)
# │   │   │   ├── Convert auth types
# │   │   │   ├── Convert auth handlers
# │   │   │   └── ...
# │   ├── Phase 2: API layer (4 subplans)
# │   └── Phase 3: Integration tests (2 subplans)

# Correct mid-level decision if needed
agents [--data-dir PATH] [--config-path PATH] plan correct <module_decision_id> --mode=revert --guidance "Use Zod instead of io-ts for validation"
# Only affected module and its subplans recompute

# Complete and apply
agents [--data-dir PATH] [--config-path PATH] plan status <plan_id>
agents [--data-dir PATH] [--config-path PATH] plan apply <plan_id>
```

**Technical Criteria:**
- Projects with 10,000+ files index without timeout.
- Context window management works (hot/warm/cold tiers).
- Hierarchical decomposition creates 4+ levels of subplans.
- Decision correction at any level recomputes only affected subtree.
- Parallel execution scales to 10+ concurrent subplans.
- Memory usage stays bounded (lazy context loading).
- A realistic porting task (500 file Python → TypeScript) completes autonomously.

**M6 SUCCESS CRITERIA** (Day 30):
- 10,000+ file project indexes with bounded memory and hot/warm/cold tiering.
- Hierarchical decomposition reaches 4+ levels with correction limited to affected subtree.
- Parallel execution scales to 10+ subplans with merge and conflict handling.
- Autonomous porting task completes with validation and review gates.
- `nox` passes with coverage >=97% including large-project suites.

---

## QUICK REFERENCE: TASK DEPENDENCIES

```
LEGEND:
  ───► = Sequential dependency (must complete before next)
  ═══► = Parallel tracks that can proceed simultaneously
  ⊕    = Merge point (all parallel tracks must complete)

WEEK 1 CRITICAL PATH:

DAY 1-2: DATA LAYER
┌─────────────────────────────────────────────────────────────────┐
│  [Jeff] A5.alpha DB migrations ═══► [Luis] A5.beta ORM models    │
│              │                           │                       │
│              └───────────⊕───────────────┘                       │
│                          │                                       │
│              [Jeff] A5.gamma Repos + Service + DI                │
│                          │                                       │
│              [Brent] A5.tests Persistence suites                 │
└─────────────────────────────────────────────────────────────────┘

DAY 1-3: RESOURCE LAYER (PARALLEL)
┌─────────────────────────────────────────────────────────────────┐
│  [Hamza] B1.core Domain Models                                  │
│              │                                                   │
│              ▼                                                   │
│  [Hamza] B2.persistence + B2.service + B3.cli                   │
└─────────────────────────────────────────────────────────────────┘

DAY 3-5: SANDBOX LAYER
┌─────────────────────────────────────────────────────────────────┐
│  [Luis] B4.sandbox strategy + manager ═══► [Hamza] git_worktree  │
│              │                           │                       │
│              └───────────⊕───────────────┘                       │
│                          │                                       │
│              [Luis] B4.sandbox copy_on_write stub                │
└─────────────────────────────────────────────────────────────────┘

DAY 4-7: ACTOR/SKILL LAYER
┌─────────────────────────────────────────────────────────────────┐
│  [Aditya] C1.schema/C1.examples Actor Schema                    │
│              │                                                   │
│              ▼                                                   │
│  [Jeff] C2.legacy Drop v2 actor configs                          │
│              │                                                   │
│              ▼                                                   │
│  [Aditya+Jeff] C2.loader/C2.compiler Actor Compiler             │
│              │                                                   │
│  [Jeff] C3.protocol/C3.context/C3.inline ═══► [Aditya] C7.mcp   │
└─────────────────────────────────────────────────────────────────┘

DAY 6-9: CHANGE TRACKING LAYER
┌─────────────────────────────────────────────────────────────────┐
│  [Luis] C5.model/C5.router Change Model + Router                │
│              │                                                   │
│              ▼                                                   │
│  [Luis] C5.diff + C6.pipeline Validation Pipeline               │
└─────────────────────────────────────────────────────────────────┘

DAY 8: M1 MERGE POINT ⊕
┌─────────────────────────────────────────────────────────────────┐
│  All tracks converge:                                           │
│  - Plan lifecycle (A) + Resources (B) + Actors (C) + Skills     │
│  - End-to-end verification of MVP criteria                      │
│  - All tests must pass                                          │
└─────────────────────────────────────────────────────────────────┘

WEEK 2: INTEGRATION + DECISIONS
┌─────────────────────────────────────────────────────────────────┐
│  [Jeff+Luis] C9 Plan-Actor Integration                          │
│              │                                                   │
│              ▼                                                   │
│  [Hamza] D1.domain/D2.service/D3.cli Decisions                  │
│              │                                                   │
│              ▼                                                   │
│  [Luis] E1 Subplan Model                                        │
└─────────────────────────────────────────────────────────────────┘

DAY 14: M3 MERGE POINT ⊕
┌─────────────────────────────────────────────────────────────────┐
│  Full plan lifecycle with actors verified                       │
│  Multi-file generation working                                  │
│  Validation pipeline operational                                │
└─────────────────────────────────────────────────────────────────┘

WEEK 3: DECISION CORRECTION + SUBPLANS
┌─────────────────────────────────────────────────────────────────┐
│  [Jeff] D4.revert/D4.append Decision Correction ─┐              │
│                                                  │              │
│  [Jeff+Luis] E2/E3/E4 Subplan Spawning + Merging ┤              │
│                                                  │              │
│  [Hamza] D5.db/D5.repo Decision Persistence ────┘              │
└─────────────────────────────────────────────────────────────────┘

DAY 21: M4 MERGE POINT ⊕
┌─────────────────────────────────────────────────────────────────┐
│  Decision tree viewing and correction working                   │
│  Can correct any decision and recompute affected work           │
└─────────────────────────────────────────────────────────────────┘

WEEK 4: SCALE + SERVER PREP
┌─────────────────────────────────────────────────────────────────┐
│  [Hamza] CTX1.index Context indexing                            │
│  [Luis] G3.semantic validation + perf tuning                    │
│  [Jeff] F0.stubs server connectivity (client-only)              │
└─────────────────────────────────────────────────────────────────┘

DAY 30: M6 TARGET ⊕
┌─────────────────────────────────────────────────────────────────┐
│  Handle 10,000+ file projects                                   │
│  Autonomous language porting with hierarchical subplans         │
│  Decision correction enables efficient iteration                │
└─────────────────────────────────────────────────────────────────┘
```

---

## SERVER CONNECTIVITY DEFERRAL NOTICE

**Server connectivity (WORKSTREAM F) is deferred beyond the 30-day timeline. The server is a separate project—this implementation covers the client only.**

The CleverAgents executable (`agents`) is purely a **client application** that can:
1. Run in stand-alone local-only mode (no server required)
2. Connect to an independently developed CleverAgents server for multi-user/collaborative features

During Days 1-30, client stub infrastructure is delivered via Section 9 (F0.stubs), covering the connect command, client interfaces, local/remote detection, and NotImplementedError stubs.

**What is NOT needed by Day 30:**
- Server implementation (the server is a separate project)
- Full client-server API implementation
- WebSocket client implementation
- Remote plan execution requests
- Authentication flow with server
- Permission system integration

The client code should be structured so that adding server connectivity later is straightforward, but no functional server communication code is required for the 30-day milestone. The server will be developed as a separate project.