2746 Commits

This Branch

This Branch

All Branches

Author	SHA1	Message	Date
freemo	642eb276a8	chore(agents): add mandatory labels to supervisor tracking issue creation ... Add label requirements to all 16 supervisor launch prompts in product-builder.md so that any tracking issues created by supervisors include the required Type/Automation, State/In Progress, and Priority/Medium labels from creation. This eliminates the persistent label compliance gap reported by the system watchdog, where supervisor-created tracking issues consistently missed required State/ and Priority/ labels. ISSUES CLOSED: #3070	2026-04-05 17:54:12 +00:00
freemo	f5d244cd37	fix(a2a): change A2aErrorDetail.code to int and map error constants to JSON-RPC 2.0 integer codes ... Per JSON-RPC 2.0 specification (Section 5.1), error codes must be integers. This commit fixes the protocol compliance defect where A2aErrorDetail.code was typed as str and error constants were string literals. Changes: - src/cleveragents/a2a/models.py: Change A2aErrorDetail.code from str to int; update field_validator to only validate 'message' (code no longer needs non-empty string check; Pydantic enforces int type) - src/cleveragents/a2a/errors.py: Change all error code constants from string literals to JSON-RPC 2.0 integer codes per docs/reference/a2a.md taxonomy: NOT_FOUND = -32001, AUTH_ERROR = -32002, FORBIDDEN = -32003, INVALID_STATE = -32004, PLAN_ERROR = -32008, CONFIGURATION_ERROR = -32009, VALIDATION_ERROR = -32602, INTERNAL_ERROR = -32603 Update map_domain_error() return type from tuple[str, str] to tuple[int, str] - features/steps/a2a_facade_steps.py: Update A2aErrorDetail construction to map symbolic string names to integer codes via _CODE_MAP - features/steps/a2a_facade_wiring_steps.py: Update error code assertion to map symbolic names to integers for comparison - features/steps/a2a_facade_coverage_boost_steps.py: Same as above - features/steps/a2a_jsonrpc_wire_format_steps.py: Update all A2aErrorDetail constructions and JSON-RPC dict payloads to use integer codes - robot/helper_a2a_facade_wiring.py: Update wired_error_mapping() to compare against integer codes - robot/helper_a2a_jsonrpc_wire_format.py: Update response_error_wire_format() to use integer code -32001 instead of string 'NOT_FOUND' Wire format now produces {"code": -32001, ...} instead of {"code": "NOT_FOUND", ...}, making it compliant with JSON-RPC 2.0 and interoperable with standards-conformant clients. ISSUES CLOSED: #2746	2026-04-05 17:53:56 +00:00
freemo	32b352832f	fix(cli): replace manual DI wiring in _get_tool_registry_service with container resolution ... _get_tool_registry_service in cli/commands/validation.py was manually constructing the full ToolRegistryService dependency graph (create_engine, sessionmaker, ToolRegistryRepository, ValidationAttachmentRepository) instead of delegating to the DI container. This duplicated wiring logic that belongs exclusively in the container and made the function harder to test. Changes: - Add _build_tool_registry_service() builder function to container.py following the established _build_skill_service/_build_session_service pattern - Register tool_registry_service as a Singleton provider in the Container class - Refactor _get_tool_registry_service() to delegate to container.tool_registry_service() — a one-liner consistent with _get_skill_service() in cli/commands/skill.py - Add TDD Behave feature (tdd_di_tool_registry_service.feature) with two scenarios: (1) function delegates to container, (2) container exposes the provider — both scenarios were failing before this fix - Update validation_cli_uncovered_branches_steps.py to match the new container-delegation pattern (mock container.tool_registry_service() instead of container.database_url()) ISSUES CLOSED: #3006	2026-04-05 17:52:59 +00:00
freemo	8d49a3b242	fix(tui): add plain text format support to session export command ... Add plain text (txt) export format to the TUI /session:export command. - Add Session.as_export_plain_text() domain method that produces a human-readable plain text transcript without Markdown formatting - Update TuiCommandRouter._session_export() to accept 'txt' as a valid format, building the plain text content via as_export_plain_text() - Update the invalid-format error message to include 'txt' as a valid option alongside 'json' and 'md' - Add BDD scenarios covering plain text export via TUI command and domain model, including with-messages and no-messages cases - Add corresponding step definitions for the new BDD scenarios The plain text format uses a simple separator-based layout: Session: <id> Actor: <actor> ... ---------------------------------------- [0] USER (timestamp): message content ---------------------------------------- ISSUES CLOSED: #3036	2026-04-05 17:52:30 +00:00
freemo	0f9ca00ce7	fix(tui): show command descriptions in SlashCommandOverlay alongside command names ... Update SlashCommandOverlay.set_commands() to accept list[SlashCommandSpec] instead of list[str], and render each entry as aligned columns: /command-name Description text Changes: - slash_command_overlay.py: update set_commands() signature to accept list[SlashCommandSpec]; render name + description with aligned padding using _COMMAND_COL_WIDTH = 28 characters for the name column - slash_catalog.py: add slash_command_specs() helper returning all specs - app.py: switch on_mount() to call slash_command_specs() instead of slash_command_names() so full spec objects are passed to the overlay - features/tui_slash_command_overlay_coverage.feature: add scenario verifying description rendering; update existing scenarios to use SlashCommandSpec objects via _make_specs() helper - features/steps/tui_slash_command_overlay_coverage_steps.py: update step implementations to build SlashCommandSpec objects from CSV names - features/tui_slash_overlay_descriptions.feature: new feature file with BDD scenarios covering description rendering and @tdd_expected_fail capture of the pre-fix name-only behaviour - features/steps/tui_slash_overlay_descriptions_steps.py: step defs for the new feature file - robot/tui_smoke.robot: add Slash Command Overlay Renders Descriptions integration test verifying descriptions appear in rendered overlay ISSUES CLOSED: #3437	2026-04-05 17:50:51 +00:00
freemo	abf250901f	fix(tool): implement tool_type filter in ToolRegistry.list_tools() ... The tool_type parameter in ToolRegistry.list_tools() was explicitly silenced with '_ = tool_type', making it a no-op. Both 'agents tool list --type validation' and 'agents tool list --type tool' returned the full unfiltered tool list. Changes: - src/cleveragents/tool/runtime.py: Add tool_type: Literal['tool', 'validation'] field to ToolSpec (default 'tool') so the registry has a discriminator to filter on. - src/cleveragents/tool/registry.py: Remove '_ = tool_type' no-op and implement actual filter: specs = [s for s in specs if s.tool_type == tool_type]. Update docstring to reflect the parameter is now active. - features/consolidated_tool.feature: Add 7 comprehensive Behave scenarios covering all filter combinations (tool, validation, None, default type, exclusion). - features/steps/tool_runtime_steps.py: Add step definition for registering a ToolSpec with an explicit tool_type. - robot/helper_tool_cli.py: Add tool_list_type_tool() and tool_list_type_validation() helper functions that verify the CLI passes tool_type through to the service layer. - robot/tool_cli.robot: Add two Robot test cases for 'tool list --type tool' and 'tool list --type validation'. The CLI handler (cli/commands/tool.py) and database repository (infrastructure/database/repositories.py) already passed tool_type through correctly — no changes needed there. ISSUES CLOSED: #2974	2026-04-05 17:49:35 +00:00
freemo	dd09fadf29	fix(exceptions): replace Any with str | os.PathLike | None for FileSystemError.path ... Narrow the path parameter type hint in FileSystemError.__init__ from Any to str | os.PathLike[str] | None. This improves static analysis accuracy since path is semantically a filesystem path, not an arbitrary value. Changes: - Add import os to exceptions.py - Replace path: Any = None with path: str | os.PathLike[str] | None = None - Reformat __init__ signature to multi-line for line-length compliance - Audit call-sites (context_service.py, project_service.py): both pass pathlib.Path which implements os.PathLike[str], no changes needed - Add BDD feature file covering str, pathlib.Path, None, and annotation check - Add corresponding step definitions ISSUES CLOSED: #3034	2026-04-05 17:49:20 +00:00
freemo	89c215e62c	fix(cli): render spec-required panels in agents plan rollback rich output ... Implemented a rich output upgrade for the agents plan rollback in the CLI by replacing the previous plain-text output in rollback_plan() with a set of four spec-required Rich panels. - Rollback Summary panel: includes Plan, Checkpoint, Label (only if available), and Files fields. - Changes Reverted panel: a Rich table with File and Action columns; supports both dict entries and plain string paths for forward compatibility with issue #2454. - Impact panel: shows Sandbox state and optional fields Child Plans Invalidated, Decisions After CP, Tool Calls After CP. - Post-Rollback State panel: includes Phase, State, and optional Checkpoints Remaining fields. - Confirmation line: "✓ OK Rollback complete" displayed after the panels. - Tests updated: Updated BDD test assertions in features/plan_cli_coverage_r2.feature to align with new panel-based output. Key design decisions: - Used getattr(result, field, None) pattern for optional fields not yet present in RollbackResult model (fields: label, child_plans_invalidated, decisions_after_cp, tool_calls_after_cp, phase, state, checkpoints_remaining) to support forward-compatibility as fields get added. - Handling mixed formats in changed_paths/changes_reverted: both dict-format entries and string-format paths are supported to remain compatible with issue #2454. - Sandbox state defaults to "restored to {checkpoint_id}" when not explicitly provided. Files changed: - src/cleveragents/cli/commands/plan.py (rollback_plan rich output section) - features/plan_cli_coverage_r2.feature (updated test assertions) ISSUES CLOSED: #2591	2026-04-05 17:35:35 +00:00
freemo	1783f0a211	docs(timeline): update schedule adherence Day 95 final v3 (2026-04-05) ... Final Day 95 timeline update (v3) with current Forgejo data as of 2026-04-05. Key Changes: - Open PRs: 92 → 170 (+78 new PRs from agent-driven activity) - Open bugs: 536 → 809 (+273 new bugs from UAT/bug-hunt agents) - Total open issues: 1547 - Session tracker: updated to #3377 - Milestone completions updated (all declined due to agent-driven issue creation outpacing closures) - Gantt chart footer, update log, and both legend blocks updated - Current Status Summary, Schedule Risk Summary, and Day 95 adherence entry all updated with final data - New UAT bugs noted: #3444 (TUI HelpPanelOverlay keybinding mismatch), #3445 (plan rollback Rich output format incomplete) Supersedes PR #3419 (docs/timeline-day95-v2) which had intermediate data. Co-authored-by: Jeffrey Phillips Freeman <the@jeffreyfreeman.me> Co-committed-by: Jeffrey Phillips Freeman <the@jeffreyfreeman.me>	2026-04-05 17:29:50 +00:00
freemo	f945e15572	fix(agents): Fix issue-implementor supervisor to properly manage N parallel workers ... The issue-implementor supervisor was defining but not using its worker dispatch logic, causing it to run only 1 worker at a time instead of the configured N parallel workers. This fix implements proper pool supervision: - Implement sliding window dispatch pattern to maintain N active workers - Use curl with prompt_async for asynchronous worker launches - Track PR workers and issue workers separately with proper monitoring - Add explicit worker count reporting in health signals (X/Y format) - Integrate PR priority gate - no new issues until all PRs have workers - Fix session monitoring and cleanup for completed/failed workers - Update product-builder heartbeat to show worker pool status The supervisor now continuously fills empty worker slots for maximum throughput, properly managing up to CA_MAX_PARALLEL_WORKERS parallel workers as designed. ISSUES CLOSED: #1	2026-04-05 16:50:47 +00:00
freemo	67b48ee817	feat!: restructure PR workflow to keep implementors accountable through merge ... BREAKING CHANGE: This completely changes how PRs are handled in the system. Implementors now own their work from creation through merge, and reviewers focus solely on code quality assessment. Major changes: - issue-implementor: Adds absolute PR prioritization - no new issues until all PRs have workers. Dispatches workers in two modes: 'pr-fix' for existing PRs and 'issue-impl' for new issues. - ca-issue-worker: Now operates in dual mode. In 'pr-fix' mode, handles review feedback, CI fixes, and merging. In 'issue-impl' mode, no longer exits after PR creation - monitors the PR until merged. - ca-continuous-pr-reviewer: Simplified to ONLY dispatch code reviewers. Removed all fix, merge, and lifecycle management. Uses dynamic review focus areas to catch different types of issues. - ca-pr-self-reviewer: Removed ALL capabilities beyond code review. No longer fixes issues, merges PRs, or manages issue states. Provides actionable feedback using rotating focus areas. - ca-pr-checker: Clarified that it should only be invoked by ca-issue-worker, not by reviewers. Benefits: - No PR backlogs (absolute priority over new issues) - Full accountability (creator owns through merge) - Better reviews (focused on quality, not mechanics) - Context preservation (no handoffs between agents) - Cleaner history (amendments instead of fix commits) This ensures implementors are accountable for their work while reviewers provide high-quality, focused code reviews without operational overhead.	2026-04-05 16:10:01 +00:00
freemo	a887712473	fix(agents): reduce health signal spam, add story point estimation, and improve supervisor monitoring ... - Health Signal Frequency: Fixed spam from ca-test-infra-improver, ca-bug-hunter, and ca-uat-tester by changing health signals from every 2-10 cycles to every 60 cycles (~10 min intervals) - Story Point Assignment: Added automatic story point estimation to ca-project-owner and ca-human-liaison during issue verification based on subtask count and complexity (XS:1, S:2, M:3, L:5, XL:8, XXL:13) - Deep Supervisor Inspection: Enhanced product-builder to check pool supervisors every 5 heartbeats for actual worker activity, detecting zombie supervisors that are running but not dispatching workers - Watchdog Integration: Added watchdog alert monitoring to product-builder that checks for critical alerts every 3 heartbeats and takes action based on severity - Alert Format Standardization: Updated ca-system-watchdog to use structured key-value alert format for easier parsing by product-builder Fixes issues with excessive Forgejo API usage, missing story point assignments during triage, and improves overall system reliability through better monitoring.	2026-04-05 15:32:04 +00:00
freemo	cce207a7bb	fix: improve agent coordination with PR prioritization and unified status tracking ... This commit addresses two critical issues in the CleverAgents autonomous system: 1. Pull Request Bottleneck: - Added PR prioritization gate to issue-implementor that checks for open PRs before taking new issues - Implementation pool now pauses new issue work when PRs need attention (failing CI, awaiting review, stale) - Re-checks PR status every 5 cycles to ensure PRs don't accumulate - Posts clear status updates explaining why new work is paused 2. Status Issue Proliferation: - Product-builder now creates ONE canonical session state issue: '[Automated] CleverAgents Build Session - <date>' - All 16 supervisors receive the session state issue number and post ALL status updates there - Removed separate tracking issue creation from ca-uat-tester and other agents - Standardized health signal format across all agents for consistent monitoring The standardized health signal format enables system-watchdog to: - Detect zombie supervisors from a single issue - Monitor active workers per pool - Track work progress across all agents - Identify stuck or inactive agents Modified agents: - issue-implementor: Added PR prioritization gate - product-builder: Single session state issue management - All pool supervisors: Updated to use session state issue - All agents: Standardized health signal format These changes ensure PRs get merged quickly and reduce issue tracker noise.	2026-04-05 14:54:22 +00:00
freemo	0d4a47f3c0	fix(config): align Settings and alembic/env.py database_url defaults to spec-required ~/.cleveragents/cleveragents.db ... - Fix Settings.database_url default from relative 'sqlite:///cleveragents.db' to absolute 'sqlite:////home/<user>/.cleveragents/cleveragents.db' using Path.home() / '.cleveragents' / 'cleveragents.db' - Fix test_database_url default similarly to use ~/.cleveragents/cleveragents_test.db - Fix alembic/env.py fallback default from Path.cwd() / '.cleveragents' / 'db.sqlite' to Path.home() / '.cleveragents' / 'cleveragents.db' (matching Settings) - Both Settings and alembic/env.py now resolve to the same absolute path - CLEVERAGENTS_DATABASE_URL env var override still works correctly in both - Update coverage_boost_steps.py assertions to match new correct defaults - Add 3 new Behave scenarios covering correct database_url default and env var override ISSUES CLOSED: #2871	2026-04-05 10:17:24 +00:00
freemo	a1fb804f21	fix(cli): correct automation-profile list output structure and rich table rendering ... The non-rich output branch of list_profiles() in src/cleveragents/cli/commands/automation_profile.py has been refactored to produce the spec-required envelope rather than a flat list of full profile objects. The new envelope structure is: {"profiles": [...simplified...], "summary": {"built_in": N, "custom": N, "total": N}} Where profiles entries now contain only five fields: name, source, select_tool, sandbox, and description. Key changes: - Profiles in the profiles array are reduced to match the exact spec (name, source, select_tool, sandbox, description). Full details are preserved for the show command via the existing _profile_spec_dict(). - The summary object aggregates counts after namespace/regex filtering so the numbers accurately reflect what is being displayed. - The rich table column header renamed from 'Select Tool' to 'Auto-Apply' per spec line 16946, and a Summary panel is appended to the rich output showing Built-in / Custom / Total counts per spec lines 16959-16963. - Behave feature file updated with new scenarios covering JSON/YAML output structure, rich table column header, and Summary panel. - New step definitions include a comprehensive profiles-wrapper-with-summary step that validates the full spec-required structure. - Robot Framework helper test_list_json() updated to validate the new dict structure with profiles wrapper and summary instead of the old flat list. ISSUES CLOSED: #2064	2026-04-05 09:46:16 +00:00
freemo	9210fbce33	refactor(cli): deduplicate session list summary logic in list_sessions ... Refactor list_sessions to reuse the summary dict already computed by _session_list_dict instead of recalculating total_msgs, sorted_sessions, most_recent, and oldest independently. The rich table summary panel now reads directly from data["summary"], ensuring JSON and rich table output are always consistent and the summary logic lives in exactly one place (_session_list_dict). Add Behave feature (session_list_summary_dedup.feature) with 6 scenarios verifying that JSON and rich table summary values are consistent across all required fields: total, most_recent, oldest, total_messages, storage. ISSUES CLOSED: #3046	2026-04-05 09:38:43 +00:00
freemo	56f424714c	fix(tdd): replace non-standard @tdd_bug tags with @tdd_issue per CONTRIBUTING.md ... Audit all feature files for non-compliant @tdd_bug tag usage and replace with the CONTRIBUTING.md-specified @tdd_issue / @tdd_issue_<N> tags. Three files were affected: - features/tdd_actor_list_no_db_update.feature: @tdd_bug @tdd_bug_797 → @tdd_issue @tdd_issue_797 (bug #797 is fixed; permanent regression guard, no @tdd_expected_fail) - features/tdd_exec_env_resolution_precedence.feature: Removed duplicate @tdd_bug @tdd_bug_1080 tags; file already had the correct @tdd_issue @tdd_issue_1080 tags. - features/tdd_use_action_automation_profile.feature: @tdd_bug @tdd_bug_1076 → @tdd_issue @tdd_issue_1076 (bug #1076 is fixed in upstream; @tdd_expected_fail not needed) Zero occurrences of @tdd_bug or @tdd_bug_<N> remain in features/. All nox quality gates pass: lint ✅ unit_tests ✅ (14426 scenarios passed). ISSUES CLOSED: #2779	2026-04-05 09:09:06 +00:00
freemo	c1dca80bf2	fix(resource-type): require letter-start for namespace/name in _NAMESPACED_RE ... Update _NAMESPACED_RE in resource_type.py and _resource_type_validation.py to require both namespace and name components to start with a letter, consistent with _BUILTIN_NAME_RE and the spec requirement that names start with a letter. Previously, _NAMESPACED_RE used [a-zA-Z0-9] as the first character class, allowing digit-starting names like '123abc/my-type'. Now uses [a-zA-Z] to enforce letter-start for both components. Also adds Behave scenarios covering rejection of digit-starting custom resource type names and valid letter-starting names. ISSUES CLOSED: #2983	2026-04-05 09:03:32 +00:00
freemo	ffb67e15b9	Merge pull request 'chore(agents): add auto-rebase on conflict to PR reviewer pool supervisor' (#1411) from improvement/pr-reviewer-auto-rebase-on-conflict into master	2026-04-05 09:01:17 +00:00
freemo	9f62f9c9c8	fix(skills): replace get_tool() calls with get() in SkillRegistry validation methods ... SkillRegistry.validate_plan() and validate_skill() called the non-existent ToolRegistry.get_tool() method, causing AttributeError at runtime. The correct method is ToolRegistry.get(), which is already used by SkillRegistry.refresh(). Changes: - Replace self._tool_registry.get_tool(entry.name) with self._tool_registry.get(entry.name) in validate_plan() - Replace self._tool_registry.get_tool(ref) with self._tool_registry.get(ref) in validate_skill() - Update Behave BDD tests to use mock.get instead of mock.get_tool - Add new BDD scenarios covering validate_skill() happy path, not-found path, unregistered include path, and no-tool-registry path - Remove stale get_tool mock setup from test helpers ISSUES CLOSED: #2914	2026-04-05 08:59:59 +00:00
freemo	c1ebacd1b9	fix(a2a): rename A2aErrorDetail.details to data per JSON-RPC 2.0 spec ... The A2aErrorDetail model used a field named 'details' for the optional error payload, but JSON-RPC 2.0 §5.1 and the project specification both require this field to be named 'data'. Changes: - Rename A2aErrorDetail.details -> data in src/cleveragents/a2a/models.py - Add TDD BDD scenarios to features/a2a_jsonrpc_wire_format.feature that verify A2aErrorDetail serializes with 'data' field (not 'details') - Add corresponding step definitions in a2a_jsonrpc_wire_format_steps.py No call sites used the 'details' keyword argument when constructing A2aErrorDetail instances (all existing usages only set code and message), so no other files required changes. ISSUES CLOSED: #2745	2026-04-05 08:57:12 +00:00
freemo	0536035d24	fix(a2a): update A2aVersionNegotiator to support JSON-RPC version 2.0 ... Implemented changes to align the A2A version negotiation with JSON-RPC 2.0 as per the specification and updated the test suite accordingly. What was implemented - Updated A2aVersionNegotiator.CURRENT_VERSION from "1.0" to "2.0" in src/cleveragents/a2a/versioning.py - Updated A2aVersionNegotiator.SUPPORTED_VERSIONS from ("1.0",) to ("2.0",) in src/cleveragents/a2a/versioning.py - Updated 7 Behave scenarios in features/consolidated_misc.feature to reflect "2.0" as the supported version: - Negotiate supported version succeeds: now negotiates "2.0" instead of "1.0" - Negotiate unsupported version raises error: now uses "99.0" as the unsupported version (since "2.0" is now supported) - is_supported returns True for valid version: now checks "2.0" instead of "1.0" - get_current returns current version: now expects "2.0" instead of "1.0" - M6 smoke A2A version negotiation accepts 2.0: updated from "1.0" to "2.0" - M6 smoke A2A version negotiation rejects unsupported: now uses "99.0" instead of "2.0" - M6 smoke A2A version is_supported returns correct result: now checks "2.0" instead of "1.0" Key design decisions - The A2A protocol is built on JSON-RPC 2.0 per the specification. The version negotiator must be consistent with JSONRPC_VERSION = "2.0" in models.py. - "1.0" is removed from SUPPORTED_VERSIONS as there is no backward compatibility requirement for the old version in the spec. - All tests updated to reflect the corrected behavior. Impacted modules/components - src/cleveragents/a2a/versioning.py - features/consolidated_misc.feature - Behavioral tests referencing A2A version negotiation ISSUES CLOSED: #2747	2026-04-05 08:52:19 +00:00
freemo	884fe12511	``` ... fix(tool): wire 6-level execution environment precedence chain into ToolRunner What was implemented - Updated ToolRunner.execute() to accept plan_priority and project_priority parameters (optional) and to propagate these priority values through the execution path. - Replaced the legacy four-level resolve_and_validate() with resolve_with_precedence() to implement the correct six-level precedence chain. - Introduced devcontainer_available derived from has_devcontainer() on linked_resource_types and wired this into the resolution process. - Updated all call sites (router.py and actor_runtime.py) to accept and forward the new priority parameters. - Added 12 unit tests verifying the six-level precedence chain behavior in ToolRunner.execute(). - Added integration tests covering override vs. fallback scenarios to ensure correct end-to-end behavior. Key design decisions and rationale - Use resolve_with_precedence() instead of resolve_with_dag() because linked_resource_types already provides devcontainer availability information, eliminating the need for a full DAG walk while preserving correct precedence semantics. - Preserve the existing contract of raising ContainerUnavailableError by retaining a final validate_container_available() call after resolution, ensuring proper error signaling when a container is resolved but no linked container resource exists. - Make plan_priority and project_priority optional with None defaulting to fallback semantics, aligning with the resolver's _parse_priority() behavior and keeping backward-compatible defaults for existing callers. ISSUES CLOSED: #2592 ```	2026-04-05 08:47:35 +00:00
freemo	faf7991a13	fix(domain): add repository protocol interfaces to domain layer ... Implemented domain-level repository protocol interfaces to restore clean architecture boundaries and enable robust type checks via runtime structural typing. - Created domain/repositories package with four protocol interfaces - Added LifecyclePlanRepositoryProtocol, ActionRepositoryProtocol, DecisionRepositoryProtocol, and ProjectRepositoryProtocol - Updated infrastructure repositories to explicitly inherit protocols - Added 10 Behave scenarios for protocol compliance verification ISSUES CLOSED: #2873	2026-04-05 08:43:01 +00:00
freemo	4d31f0ed02	fix(cli): promote --format to global CLI callback option per spec ... Add `fmt: OutputFormat` parameter to `main_callback()` in `src/cleveragents/cli/main.py` and store the selected format in `ctx.obj["format"]` so all subcommands can read it without needing their own per-command `--format` flag. Remove per-command `--format` / `fmt` parameters from `version()`, `info()`, and `diagnostics()` commands. These commands now read the format from `ctx.obj.get("format", OutputFormat.RICH.value)`. The specification states: "The framework supports six distinct output formats, selectable via the global `--format` flag." This change aligns the implementation with the spec by making `--format` a global option on the root `agents` command (via the Typer callback). All six formats (json, yaml, plain, rich, table, color) are supported via the global flag and the `-f` shorthand. Add Behave BDD scenarios covering global `--format` flag propagation to subcommands for all six formats. Update Robot Framework integration tests to exercise the global `--format` flag. Update existing tests that used per-command `--format` for version/info/diagnostics to use the global flag instead. ISSUES CLOSED: #2908	2026-04-05 08:39:14 +00:00
freemo	fc7e47f66b	fix(cli): remove --namespace/-n from agents plan list command ... Remove the --namespace/-n option from the agents plan list CLI command as it is not defined in the specification. The spec defines: agents plan list [--phase <PHASE>] [--state <STATE>] [--project <PROJECT>] [--action <ACTION>] [<REGEX>] Namespace filtering is already supported via the positional <REGEX> argument (e.g. 'agents plan list "^myteam/"'). Changes: - Remove namespace parameter from lifecycle_list_plans() function - Remove namespace=namespace from service.list_plans() call - Remove Namespace row from TUI Filters panel - Update docstring examples to remove --namespace references - Update filter panel condition to exclude namespace check - Remove 4 namespace-related scenarios from plan_cli_spec_alignment.feature - Remove corresponding step definitions from plan_cli_spec_alignment_steps.py ISSUES CLOSED: #2986	2026-04-05 08:38:26 +00:00
freemo	4b9798bc40	fix(sandbox): use spec-required cleveragents/plan- branch prefix in GitWorktreeSandbox ... - Implemented branch naming change in GitWorktreeSandbox.create() from sandbox/<plan_id> to cleveragents/plan-<plan_id>, located in src/cleveragents/infrastructure/sandbox/git_worktree.py - Updated the create() method docstring to reflect the new branch naming convention - Updated test fixtures: - features/steps/git_worktree_coverage_boost_steps.py: sandbox/test -> cleveragents/plan-test, sandbox/cleanup-test -> cleveragents/plan-cleanup-test, sandbox/branch-delete-fail -> cleveragents/plan-branch-delete-fail - Updated scenario: - features/garbage_collection.feature: sandbox/test-plan -> cleveragents/plan-test-plan - Design rationale: - Adopted the spec-required format: cleveragents/plan-{safe_plan_id} to align with the spec-defined namespace for all CleverAgents-managed git branches - Ensures consistent branch naming across code, tests, and features for easier policy enforcement and maintainability ISSUES CLOSED: #2601	2026-04-05 08:29:55 +00:00
freemo	21a551cd6f	fix(resource): use namespace column instead of name heuristic in db_to_spec() for built_in field ... Replace the name-based heuristic ("/" not in name) in db_to_spec() with the authoritative namespace column check (str(raw_ns) == "builtin"), consistent with the existing _load_type_registry() implementation. The old heuristic would misclassify custom resource types whose names contain no "/" as built-in, potentially blocking their removal via remove_type(). The namespace column is set to "builtin" by spec_to_db() for all built-in types and to the actual namespace prefix for custom types, making it the correct source of truth. A name-heuristic fallback is retained for legacy rows where the namespace column is NULL (e.g. rows inserted directly without going through spec_to_db()). New BDD scenarios added to resource_registry_service_coverage.feature: - _db_to_spec sets built_in True when namespace column is "builtin" - _db_to_spec sets built_in False when namespace column is not "builtin" ISSUES CLOSED: #3013	2026-04-05 08:29:36 +00:00
freemo	bb97f87a31	docs(spec): fix invariant precedence chain (3-tier->4-tier) and document non_overridable flag ... Approved proposal: #3064 Changes: - Glossary (line 92): Fix 3-tier precedence claim to correct 4-tier chain (plan > action > project > global) and add non_overridable exception note - Line 18980: Fix Invariant Reconciliation Actor step 5 precedence reference - Line 18991: Fix Strategize phase invariant reconciliation precedence reference - Lines 19601-19604: Update Precedence and conflict resolution section to include action tier as a distinct scope (not promoted to plan tier) and add non_overridable exception clause - Line 19618: Fix invariant view calculation precedence reference - Lines 19621+: Add Non-overridable global invariants documentation block Rationale: The glossary and two Strategize-phase descriptions incorrectly stated a 3-tier precedence chain (plan > project > global), contradicting ADR-016 which defines a 4-tier chain (plan > action > project > global). The implementation in actor/reconciliation.py and application/services/plan_lifecycle_service.py correctly uses the 4-tier model. The non_overridable field on the Invariant domain model was implemented but completely undocumented in the spec.	2026-04-05 08:27:30 +00:00
freemo	cd162488ec	fix(resource-registry): implement get_parents() on ResourceDagMixin ... Add get_parents(name_or_id: str) -> list[Resource] method to ResourceDagMixin in _resource_registry_dag.py. The method is symmetric to the existing get_children() method but queries ResourceLinkModel by child_id instead of parent_id to return all direct parent resources. Implementation details: - Calls show_resource() first to raise NotFoundError for non-existent resources - Queries ResourceLinkModel filtered by child_id=resource.resource_id - Returns deterministically sorted list (by name, then resource_id) - Exposed via ResourceRegistryService through the mixin inheritance chain The BDD scenario 'Get parents returns all direct parents' in resource_dag.feature now passes. All 15 existing @dag_traversal and other DAG scenarios continue to pass with no regression. ISSUES CLOSED: #2844	2026-04-05 08:27:19 +00:00
freemo	6e09842731	fix(cli): raise on subcommand registration failure instead of partial silent return ... - What was implemented - Fixed _register_subcommands in src/cleveragents/cli/main.py: replaced bare return with raise SystemExit(1) from exc in the exception handler to propagate a non-zero exit code and preserve the original traceback context. - Removed # pragma: no cover from the exception handler, enabling test coverage for this path. - Why this change - Ensures the CLI exits with a non-zero status on subcommand registration failure and provides proper error context for debugging, aligning with the project's error-handling goals and B904 linting requirements. - Tests and verification - Added Behave BDD issue-capture test features/tdd_cli_incomplete_subcommand_registration.feature with 3 scenarios to exercise and validate the failure path. - Added step definitions features/steps/tdd_cli_incomplete_subcommand_registration_steps.py corresponding to the new scenarios. - The @tdd_expected_fail scenario captures the previous buggy behavior (silent return with exit code 0) to ensure regression is addressed. - All nox quality gates pass (lint, typecheck). - Key design decisions - Use raise SystemExit(1) from exc to exit with a clear non-zero status while preserving the original exception chain (satisfies B904). - Coverage ensured by removing the pragma, bringing the error path under test. - Behavior now explicitly signals failure to the shell and any orchestrating tooling, avoiding silent failures. - Affected modules and artifacts - src/cleveragents/cli/main.py - features/tdd_cli_incomplete_subcommand_registration.feature - features/steps/tdd_cli_incomplete_subcommand_registration_steps.py ISSUES CLOSED: #2604	2026-04-05 08:27:05 +00:00
freemo	e1f5c95bad	fix(database): move get_all_for_project call outside loop in LegacyDataMigrator ... Move ctx.plans.get_all_for_project(project.id) outside the plan iteration loop in LegacyDataMigrator.migrate_project_data to eliminate an N+1 database query pattern. Previously, the method fetched all plans from the database on every iteration of the plans loop, resulting in O(N) queries where N is the number of plans in the legacy plans.json file. For projects with many plans, this caused significant unnecessary database load during migration. The fix fetches all existing plans once before the loop begins, then uses the in-memory list for duplicate detection on each iteration. Also verified no other similar N+1 patterns exist in LegacyDataMigrator. Added a new Behave scenario 'get_all_for_project is called only once for multiple plans' that patches the repository method to count invocations and asserts exactly one call regardless of how many plans are migrated. ISSUES CLOSED: #3047	2026-04-05 08:26:50 +00:00
freemo	9664271562	fix(resources): allow agents resource stop to stop container-instance and devcontainer-instance resources ... Add container-instance to _STOPPABLE_TYPES so that agents resource stop accepts both container-instance and devcontainer-instance resources, as required by the specification. The stop_container function already uses docker stop with the container_id from the lifecycle tracker, which works generically for both resource types. - Add container-instance to _STOPPABLE_TYPES frozenset - Update resource_stop() docstring to reflect both stoppable types - Update devcontainer_cleanup.feature: F19 scenario now expects success for container-instance stop (was incorrectly expecting rejection) - Add TDD feature file and step definitions for issue #2588 regression guard ISSUES CLOSED: #2588	2026-04-05 08:22:45 +00:00
freemo	e05a577090	fix(acms): align DEFAULT_SKELETON_RATIO default value with spec ... Align all DEFAULT_SKELETON_RATIO / _DEFAULT_SKELETON_RATIO constants to the spec-required value of 0.15 (docs/specification.md line 35294). Previously three modules used divergent defaults: - skeleton_compressor.py: 0.3 (2x the spec value) - depth_breadth_projection.py: 0.2 (33% above spec) - project_context.py: 0.2 (33% above spec) This caused child plans to receive 2-3x more skeleton context than intended, and produced inconsistent behaviour depending on whether skeleton compression was invoked via the CLI or the service layer. Changes: - Set DEFAULT_SKELETON_RATIO = 0.15 in skeleton_compressor.py - Set DEFAULT_SKELETON_RATIO = 0.15 in depth_breadth_projection.py - Set _DEFAULT_SKELETON_RATIO = 0.15 in project_context.py - Update Behave feature tests to assert the 0.15 default value - Add new scenario to project_context_cov3.feature asserting _DEFAULT_SKELETON_RATIO == 0.15 in _default_acms_config() ISSUES CLOSED: #2909	2026-04-05 08:22:36 +00:00
freemo	2d07cd5ef8	fix(cli): add NAME positional argument to agents actor add command per spec ... The spec (docs/reference/actor_cli.md) defines the synopsis for `agents actor add` as: agents actor add <NAME> --config <FILE> [--update] [--unsafe] [--set-default] [--option key=value] [--format FORMAT] The implementation was missing the required <NAME> positional argument, silently reading the actor name from the config file's `name` field instead. This deviates from the spec and breaks the expected CLI UX. Changes: - Add `name` as a required positional Argument to the `add` command - Update docstring to match spec synopsis exactly - The positional NAME takes precedence over any `name` field in config - Remove the now-redundant config-file name validation (name comes from CLI) - Add Behave BDD feature + steps for the NAME positional argument (TDD) - Update all existing Behave step invocations to pass NAME positional arg - Update Robot Framework helpers to pass NAME positional arg ISSUES CLOSED: #2905	2026-04-05 08:12:32 +00:00
freemo	8d7f2103e7	fix(resource-registry): remove type: ignore[override] suppressions from ResourceDagMixin by restructuring mixin/protocol relationship ... The two # type: ignore[override] inline suppression comments on the link_child and unlink_child methods of ResourceDagMixin violated the project's strict no-suppression policy in CONTRIBUTING.md. Investigation revealed the suppressions were unnecessary: after removing them, nox -e typecheck passes with 0 errors and 0 warnings. The underlying Pyright type checker already accepts the mixin pattern with self: RegistryHost annotations without any suppression. Both comments are removed with no other changes required. ISSUES CLOSED: #2830	2026-04-05 08:07:21 +00:00
freemo	dbf3b8d7f7	fix(resources): remove overlay from SandboxStrategy enum - not in spec ... Add 'overlay' to _VALID_STRATEGIES frozenset in ResourceTypeConfigSchema validator so that resource type YAML configs specifying sandbox_strategy: overlay are accepted without raising ValueError. The overlay strategy is defined in the specification for fs-mount resources using OverlayFS. The domain model SandboxStrategy enum correctly includes OVERLAY = 'overlay', but the YAML schema validator in src/cleveragents/resource/schema.py was missing it from _VALID_STRATEGIES, causing a validation error on valid configurations. Changes: - Add 'overlay' to _VALID_STRATEGIES in src/cleveragents/resource/schema.py - Add BDD scenario: 'Schema accepts overlay sandbox strategy' - Add step definitions for the new scenario ISSUES CLOSED: #2827	2026-04-05 08:04:27 +00:00
freemo	03e5403374	chore(agents): add auto-rebase on conflict to PR reviewer pool ... Agent evolver identified a systematic pattern: - Pattern: Dead-end conflict handling in PR reviewer - Evidence: When the reviewer detects merge conflicts, it posts a comment saying 'implementor needs to rebase' and marks the PR as done. But the issue worker has already exited after PR creation — nobody acts on the rebase request. This created a dead end where 11+ approved PRs were abandoned due to conflicts (PRs #1219, #1236, #1247, #1248, #1220, #1237, #1238, #1246, #1252, #1269). - Fix: When the reviewer reports a conflict, the pool supervisor (which has full bash permissions and maintains a clone) now attempts to rebase the PR branch onto latest master itself. If the rebase succeeds, the PR is re-queued for merge. If it fails, the PR is abandoned with a clear comment explaining manual intervention is needed. This change requires human approval before taking effect.	2026-04-05 07:53:08 +00:00
freemo	1411adfed3	Merge pull request 'fix(lsp): release lock before blocking I/O in LspLifecycleManager.restart_server() to prevent deadlock' (#3165) from fix/lsp-lifecycle-restart-lock-deadlock into master	2026-04-05 07:50:09 +00:00
freemo	491781714f	fix(actors): enforce --update flag in agents actor add - reject re-adding existing actor without --update ... Before calling upsert_actor(), check if the actor already exists using registry.get_actor() or service.get_actor(). If the actor exists and update_existing is False, print the spec-required error panel and exit with error code 1. Error panel includes: - Actor name - Registration timestamp (formatted as YYYY-MM-DD HH:MM) - Hint to use --update flag Adds Behave tests for: - Duplicate-without-update failure case (exit code 1, error panel shown) - Duplicate-with-update success case (exit code 0, actor updated) - New actor without --update success case (exit code 0, actor added) ISSUES CLOSED: #2609	2026-04-05 07:49:49 +00:00
freemo	7c2f5a1c76	fix(mcp): correct MCPToolResult.data type annotation for MCP 1.4.0 content list format ... - What was implemented - Updated MCPToolResult.data docstring to document the normalisation behaviour, clarifying that MCP 1.4.0 list-format content is normalised to a dict with a "content" key. - Updated MCPToolAdapter.invoke() success path to normalise MCP 1.4.0 list-format content: when content is a list (as per MCP 1.4.0 spec), it is wrapped as {"content": [...]} so MCPToolResult.data is always dict[str, Any]. - Updated MockMCPTransport.call() to return MCP 1.4.0-compliant list-format content: invoke_results are now wrapped as [{"type": "text", "text": json.dumps(payload)}] instead of the non-standard dict format. - Added json import to mock_mcp_transport.py to support JSON payload encoding. - Updated existing mcp_adapter.feature scenario "Invoke a discovered tool successfully" to check for "content" key instead of "id" (since mock now returns MCP 1.4.0 list format). - Added two new Behave scenarios: "Invoke tool returns MCP 1.4.0 list-format content normalised to dict" and "Invoke tool with MCP 1.4.0 content list stores content items". - Added corresponding step definitions for the new scenarios. - All 14,418 existing scenarios continue to pass; 2 new scenarios added; typecheck passes with 0 errors; lint passes. - Why this approach - Design decision: Normalize to dict (Option B) to maintain a consistent MCPToolResult.data type of dict[str, Any] and avoid breaking downstream code that accesses result.data["key"]. - Fallback path handles non-standard server responses gracefully, ensuring robustness when servers deviate from MCP 1.4.0 spec. - Technical approach and affected components - Core: MCPToolResult data handling and MCPToolAdapter.invoke() logic - Mocks: mock_mcp_transport.py updated to emit MCP 1.4.0 list-format content - Tests: updated mcp_adapter.feature expectations; added two Behave scenarios with new step definitions - Dependencies: added import json to mock_mcp_transport.py - Verification - Comprehensive test suites: 14,418 existing scenarios pass - 2 new scenarios added and pass - Typechecking: 0 errors - Linting: passes ISSUES CLOSED: #2743	2026-04-05 07:48:02 +00:00
freemo	7b0533b34d	fix(skills): add lowercase-only namespace/name pattern validation to SkillConfigSchema.name field ... Enforce lowercase-only namespace/name pattern in both SkillConfigSchema and ActionConfigSchema by updating NAMESPACED_NAME_RE from the permissive [a-zA-Z0-9] pattern to the strict [a-z0-9] pattern. This fixes a UAT-identified bug where uppercase letters were incorrectly accepted in the namespace/name fields of skill and action configurations, violating the spec's naming convention (^[a-z0-9_-]+/[a-z0-9_-]+$). Changes: - src/cleveragents/skills/schema.py: Update NAMESPACED_NAME_RE to reject uppercase letters in namespace and name parts - src/cleveragents/action/schema.py: Apply the same lowercase-only fix - features/skill_schema.feature: Add BDD scenarios for uppercase namespace, uppercase name part, fully uppercase name, uppercase tool ref, and uppercase include name rejection - features/consolidated_action.feature: Add BDD scenarios for uppercase namespace, uppercase name part, and fully uppercase name rejection ISSUES CLOSED: #3029	2026-04-05 07:43:56 +00:00
freemo	2eaf9362a3	fix(tests): add missing Behave feature file for coverage_threshold ... - What was implemented - Created features/coverage_threshold_config.feature containing 11 BDD scenarios to exercise coverage threshold configurations. - All 15 step functions in features/steps/coverage_threshold_config_steps.py are now exercised by the new scenarios. - Scenarios cover: - pyproject.toml coverage config (7 scenarios) - noxfile.py coverage session (2 scenarios) - CI workflow (1 scenario) - nightly quality workflow (1 scenario) - All 11 scenarios verified to pass against actual project configuration values. - Key design decisions - Used the exact step text from the step file to ensure proper Gherkin-to-step matching. - Nightly workflow threshold set to 85 (matching the actual --coverage-min 85 value in nightly-quality.yml). - Noxfile threshold set to 97 (matching COVERAGE_THRESHOLD = 97 constant). - Organized scenarios into logical groups with comments for readability. - Modules/components affected - features/coverage_threshold_config.feature (new feature file) - features/steps/coverage_threshold_config_steps.py (all steps now exercised) ISSUES CLOSED: #2767	2026-04-05 07:43:10 +00:00
freemo	e0336379f5	fix(robot): correct dedent logic for try...except...finally in indentation_library ... Rewrites the block_stack management in fix_python_indentation to correctly handle try...except...finally blocks in all configurations: - Single except clause - Multiple except clauses for the same try - try...except...finally - Nested try blocks with outer except - Nested try with inner finally followed by outer except (critical regression) Root cause: The original implementation stored only keyword strings in block_stack (e.g. 'try', 'class', 'other'). Two bugs existed: 1. When 'finally:' was encountered, the code popped the 'try' from the stack. This meant subsequent 'except:' clauses for an outer try could not find their matching try, producing syntactically invalid Python. 2. The dedent calculation for 'except'/'finally' counted all stack entries up to the first 'try', but did not account for nested try blocks where the inner try's except/finally had already been processed. Fix: Replace the flat string stack with a tuple stack of (keyword, base_indent, has_seen_except_finally). The new logic: - 'finally:' always belongs to the innermost try. Pop non-try entries, set indent to that try's base_indent, then pop the try itself (finally closes the block). The outer try remains on the stack. - 'except:' belongs to the innermost try. If the innermost try has already seen an except/finally AND there is an outer try, the inner try is done: pop it and use the outer try instead. Otherwise use the innermost try and mark it as having seen an except/finally clause. Adds a Behave BDD feature (tdd_indentation_library_try_except.feature) with 6 scenarios covering all acceptance criteria from issue #2845. ISSUES CLOSED: #2845	2026-04-05 07:39:57 +00:00
freemo	e4966021e9	fix(config): correct Settings.data_dir default from Path("data") to Path.home() / ".cleveragents" ... The Settings class had data_dir defaulting to Path("data") — a relative path — but the specification requires the default data directory to be ~/.cleveragents. This inconsistency caused any code reading Settings().data_dir without setting CLEVERAGENTS_DATA_DIR to use a relative 'data/' directory instead of the spec-required ~/.cleveragents, affecting log storage, database location, cache, backups, and all persistent state. Changes: - Fix data_dir default_factory from Path("data") to Path.home() / ".cleveragents" - Add Behave scenario: 'data_dir default is the spec-required home directory path' - Add Behave scenario: 'data_dir env var override takes precedence over default' - Add step definition: 'the data directory should equal the home cleveragents path' Settings.data_dir is now consistent with ConfigService core.data-dir default (~/.cleveragents). The CLEVERAGENTS_DATA_DIR env var override continues to work. ISSUES CLOSED: #2851	2026-04-05 07:33:40 +00:00
freemo	f373e3a527	fix(resources): register fs-mount built-in resource type at startup ... Add fs-mount ResourceTypeDefinition to _GIT_FS_CONTAINER_TYPES in _resource_registry_data.py so that bootstrap_builtin_types() seeds it into the registry at startup. The fs-mount type was listed in BUILTIN_TYPE_NAMES (validation frozenset) but had no corresponding type definition in BUILTIN_TYPES, causing: - 'agents resource type list' to omit fs-mount - 'agents resource add fs-mount ...' to fail with 'Resource type not found' - fs-directory to be unable to reference fs-mount as a parent type Changes: - Add fs-mount definition to _GIT_FS_CONTAINER_TYPES with correct fields: name, description, resource_kind=physical, sandbox_strategy=copy_on_write, user_addable=True, cli_args=[path], parent_types=[], child_types=[fs-directory], auto_discovery with fs-directory rule, handler=fs_mount:FsMountHandler, capabilities (read/write/sandbox=True, checkpoint=False) - Add fs-mount to fs-directory's parent_types so fs-directory can be nested under an fs-mount resource - Add BDD scenarios in resource_type_bootstrap_fs_mount.feature covering registration, user_addable, child_types, parent_types, resource add, and fs-directory parent resolution - Add Robot Framework integration tests in resource_type_bootstrap_fs_mount.robot covering all four acceptance criteria end-to-end ISSUES CLOSED: #2911	2026-04-05 07:29:35 +00:00
freemo	fc00834bd3	fix(cli): add missing _log.debug call to session export/import/tell DatabaseError handlers ... Add _log.debug("session export failed", exc_info=True), _log.debug("session import failed", exc_info=True), and _log.debug("session tell failed", exc_info=True) to the DatabaseError handlers in export_session, import_session, and tell commands respectively. This brings these handlers into consistency with the existing create, list, show, and delete handlers which already include the _log.debug call for diagnostic tracebacks when --log-level debug is used. Also adds three new BDD scenarios and step implementations that verify _log.debug is called in each of the three handlers, ensuring the pattern is tested and cannot regress. ISSUES CLOSED: #2788	2026-04-05 07:28:12 +00:00
freemo	3d8d8058c9	chore: remove stray 2n file accidentally committed	2026-04-05 07:18:40 +00:00
freemo	a5d6f0c393	fix(a2a): add A2A Python SDK as project dependency ... Add a2a-sdk>=0.3.0 to [project.dependencies] in pyproject.toml. The specification mandates that both local (stdio) and server (HTTP) transports use the A2A Python SDK (ADR-047). Previously the package was absent from pyproject.toml and uv.lock, making the project non-compliant with the spec requirement. Changes: - Add 'a2a-sdk>=0.3.0' to [project.dependencies] in pyproject.toml - Regenerate uv.lock to include a2a-sdk 0.3.25 and its transitive deps - Add Behave scenarios confirming a2a is importable as a project dependency - Add step definitions for the new TDD scenarios ISSUES CLOSED: #2922	2026-04-05 07:18:27 +00:00
freemo	da0c5f14b5	fix(config): add server_url, server_token, format, and default_estimation_actor fields to Settings ... Add four spec-required environment variable fields to the Settings class that were previously missing, causing AttributeError when accessed: - server_url: str | None — maps to CLEVERAGENTS_SERVER_URL (spec: server.url) - server_token: str | None — maps to CLEVERAGENTS_SERVER_TOKEN (spec: server.token) - format: str | None — maps to CLEVERAGENTS_FORMAT (spec: core.format) - default_estimation_actor: str | None — maps to CLEVERAGENTS_DEFAULT_ESTIMATION_ACTOR (spec: core.default_estimation_actor) All fields default to None when the corresponding env var is unset, and are fully statically typed (no type: ignore). BDD scenarios cover env var binding, default value, type validation, and model_fields presence for each new field. ISSUES CLOSED: #2866	2026-04-05 07:15:48 +00:00