cleveragents/cleveragents-core

Fork 3

[CA-AUTO] UAT Pool Supervisor — v3.6.0 — Session Tracker #2360

New issue

Open

opened 2026-04-03 17:19:16 +00:00 by freemo · 15 comments

freemo commented

2026-04-03 17:19:16 +00:00

Owner

This issue tracks the UAT pool supervisor session for milestone v3.6.0.
All progress reports will be posted as comments here.

Instance ID: uat-pool-1
Started: 2026-04-03
Master HEAD: 8866c58bd4

Feature areas under test:

Plan Lifecycle (Action/Strategize/Execute/Apply phases)
Projects & Resources
Actor System
Tool Sources & Security
Skills System
Context Management (ACMS/UKO)
API Endpoints
CLI Commands
Data Models & Schemas
Configuration & Environment
Testing Infrastructure
Sandbox & Checkpoint Model

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: ca-uat-tester

This issue tracks the UAT pool supervisor session for milestone v3.6.0. All progress reports will be posted as comments here. Instance ID: uat-pool-1 Started: 2026-04-03 Master HEAD: 8866c58bd49d8d4b9b53c5438d7f4f6e454eb740 Feature areas under test: - Plan Lifecycle (Action/Strategize/Execute/Apply phases) - Projects & Resources - Actor System - Tool Sources & Security - Skills System - Context Management (ACMS/UKO) - API Endpoints - CLI Commands - Data Models & Schemas - Configuration & Environment - Testing Infrastructure - Sandbox & Checkpoint Model --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: ca-uat-tester

freemo commented

2026-04-03 17:20:36 +00:00

Author

Owner

UAT Pool Supervisor — Initial Dispatch (Cycle 1)

Time: 2026-04-03 UTC
HEAD: 8866c58bd4

Workers Dispatched (10/10 slots filled)

Session ID	Feature Area
ses_2aba48266ffeUkX2ThbBrdL0RX	plan-lifecycle
ses_2aba47eaeffeBGrz4sDDwFTFi0	projects-and-resources
ses_2aba47b28ffeGuDOO3QPeHtf8m	actor-system
ses_2aba475ceffebf0G2JGXmamQ8K	tool-sources-and-security
ses_2aba47195ffeXBwLPvRw587hbG	skills-system
ses_2aba46eccffeJWZHDBOONksK2y	context-management
ses_2aba46b44ffebh1Nfpr3YELj39	api-endpoints
ses_2aba467bbffe46WVyrQJMUacmQ	cli-commands
ses_2aba463fcffexweqFuQKZpwIDD	data-models-and-schemas
ses_2aba460b8ffeuGQUxrwRu1lwgQ	configuration-and-environment

Remaining Areas (queued for batch 2)

sandbox-and-checkpoint-model
testing-infrastructure

Status

Active workers: 10/10
Tested areas: 0/12
Coverage: 0%
Bugs filed: 0 (so far)

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: ca-uat-tester

## UAT Pool Supervisor — Initial Dispatch (Cycle 1) **Time**: 2026-04-03 UTC **HEAD**: 8866c58bd49d ### Workers Dispatched (10/10 slots filled) | Session ID | Feature Area | |---|---| | ses_2aba48266ffeUkX2ThbBrdL0RX | plan-lifecycle | | ses_2aba47eaeffeBGrz4sDDwFTFi0 | projects-and-resources | | ses_2aba47b28ffeGuDOO3QPeHtf8m | actor-system | | ses_2aba475ceffebf0G2JGXmamQ8K | tool-sources-and-security | | ses_2aba47195ffeXBwLPvRw587hbG | skills-system | | ses_2aba46eccffeJWZHDBOONksK2y | context-management | | ses_2aba46b44ffebh1Nfpr3YELj39 | api-endpoints | | ses_2aba467bbffe46WVyrQJMUacmQ | cli-commands | | ses_2aba463fcffexweqFuQKZpwIDD | data-models-and-schemas | | ses_2aba460b8ffeuGQUxrwRu1lwgQ | configuration-and-environment | ### Remaining Areas (queued for batch 2) - sandbox-and-checkpoint-model - testing-infrastructure ### Status - Active workers: 10/10 - Tested areas: 0/12 - Coverage: 0% - Bugs filed: 0 (so far) --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: ca-uat-tester

freemo commented

2026-04-03 17:20:56 +00:00

Author

Owner

UAT Worker Starting — Tool Sources & Security

Instance ID: uat-tester-tool-sources-security
Feature Area: Tool Sources & Security
Started: 2026-04-03

Beginning analysis of tool registration, tool source management, security policies, sandboxing, and access control features.

Will post findings as bugs are discovered.

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: ca-uat-tester

## UAT Worker Starting — Tool Sources & Security **Instance ID**: uat-tester-tool-sources-security **Feature Area**: Tool Sources & Security **Started**: 2026-04-03 Beginning analysis of tool registration, tool source management, security policies, sandboxing, and access control features. Will post findings as bugs are discovered. --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: ca-uat-tester

freemo commented

2026-04-03 17:21:11 +00:00

Author

Owner

UAT Worker Starting — Data Models and Schemas

Instance: uat-tester-data-models-schemas
Feature Area: data-models-and-schemas
Working Directory: /app (read-only analysis mode)
Started: 2026-04-03

Beginning code-level analysis of all data models, Pydantic schemas, database models, and serialization/deserialization against the specification.

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: ca-uat-tester

## UAT Worker Starting — Data Models and Schemas **Instance**: uat-tester-data-models-schemas **Feature Area**: data-models-and-schemas **Working Directory**: /app (read-only analysis mode) **Started**: 2026-04-03 Beginning code-level analysis of all data models, Pydantic schemas, database models, and serialization/deserialization against the specification. --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: ca-uat-tester

freemo commented

2026-04-03 17:21:19 +00:00

Author

Owner

UAT Worker Starting — API Endpoints Feature Area

Instance: uat-tester-api-endpoints
Focus Area: api-endpoints (REST API / A2A JSON-RPC endpoints)
Mode: Worker

Plan

The specification uses A2A Protocol (JSON-RPC 2.0) for all client-server communication. I will:

Survey the codebase for API endpoint implementations
Compare against spec-defined methods
Test endpoint existence, parameter handling, response formats, and error handling
File bugs for any discrepancies

Starting code analysis now...

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: ca-uat-tester

## UAT Worker Starting — API Endpoints Feature Area **Instance**: uat-tester-api-endpoints **Focus Area**: api-endpoints (REST API / A2A JSON-RPC endpoints) **Mode**: Worker ### Plan The specification uses A2A Protocol (JSON-RPC 2.0) for all client-server communication. I will: 1. Survey the codebase for API endpoint implementations 2. Compare against spec-defined methods 3. Test endpoint existence, parameter handling, response formats, and error handling 4. File bugs for any discrepancies Starting code analysis now... --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: ca-uat-tester

freemo commented

2026-04-03 17:21:26 +00:00

Author

Owner

UAT Worker Starting — Context Management (ACMS/UKO)

Instance: uat-context-mgmt-worker
Feature Area: Context Management (ACMS + UKO)
Started: 2026-04-03

Scope

Testing the following sub-areas:

Universal Knowledge Ontology (UKO) — Layers 0-3, vocabulary registry, detail level maps
ACMS Pipeline — 10-stage, 3-phase context assembly
Context Request Protocol (CRP)
Tiered Storage (hot/warm/cold)
Skeleton Compressor (inherited context)
UKO Indexer and Query Interface
CLI context commands (agents project context set)

Method

Code-level analysis against specification requirements.

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: ca-uat-tester

## UAT Worker Starting — Context Management (ACMS/UKO) **Instance**: uat-context-mgmt-worker **Feature Area**: Context Management (ACMS + UKO) **Started**: 2026-04-03 ### Scope Testing the following sub-areas: - Universal Knowledge Ontology (UKO) — Layers 0-3, vocabulary registry, detail level maps - ACMS Pipeline — 10-stage, 3-phase context assembly - Context Request Protocol (CRP) - Tiered Storage (hot/warm/cold) - Skeleton Compressor (inherited context) - UKO Indexer and Query Interface - CLI context commands (`agents project context set`) ### Method Code-level analysis against specification requirements. --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: ca-uat-tester

freemo commented

2026-04-03 17:22:39 +00:00

Author

Owner

UAT Worker Starting — Plan Lifecycle Feature Area

Instance: uat-tester-plan-lifecycle
Focus Area: Plan Lifecycle (Action/Strategize/Execute/Apply phases)
Started: 2026-04-03

Code Analysis Summary

Completed deep code analysis of:

src/cleveragents/domain/models/core/plan.py — Plan domain model, PlanPhase, ProcessingState, VALID_PHASE_TRANSITIONS
src/cleveragents/action/schema.py — ActionConfigSchema
src/cleveragents/domain/models/core/action.py — Action domain model
src/cleveragents/application/services/plan_lifecycle_service.py — PlanLifecycleService
src/cleveragents/cli/commands/plan.py — Plan CLI commands
docs/specification.md — Plan lifecycle specification

Issues Found (filing now)

effective_profile_snapshot defaults to {} instead of capturing the actual resolved profile — spec requires a frozen JSON snapshot at plan creation time for audit/reproducibility; the code acknowledges this in a comment but doesn't fix it
agents plan list is missing --namespace/-n filter — spec requires agents plan list [--namespace NS]; the CLI command does not expose this option
VALID_PHASE_TRANSITIONS allows STRATEGIZE → EXECUTE but not STRATEGIZE → STRATEGIZE — the spec says Execute can revert to Strategize; the code correctly models this, but the can_revert_to() method on Plan uses can_transition() which doesn't allow STRATEGIZE → STRATEGIZE (needed for re-running Strategize after a revert from Execute)
Invariant precedence: spec says plan > project > global (three-tier), but code models plan > action > project > global (four-tier) — the spec glossary says "Action invariants are promoted to plan-level when the action is used, so the runtime precedence chain is three-tier: plan > project > global" — the code keeps action as a separate tier rather than promoting to plan level

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: ca-uat-tester

## UAT Worker Starting — Plan Lifecycle Feature Area **Instance**: uat-tester-plan-lifecycle **Focus Area**: Plan Lifecycle (Action/Strategize/Execute/Apply phases) **Started**: 2026-04-03 ### Code Analysis Summary Completed deep code analysis of: - `src/cleveragents/domain/models/core/plan.py` — Plan domain model, PlanPhase, ProcessingState, VALID_PHASE_TRANSITIONS - `src/cleveragents/action/schema.py` — ActionConfigSchema - `src/cleveragents/domain/models/core/action.py` — Action domain model - `src/cleveragents/application/services/plan_lifecycle_service.py` — PlanLifecycleService - `src/cleveragents/cli/commands/plan.py` — Plan CLI commands - `docs/specification.md` — Plan lifecycle specification ### Issues Found (filing now) 1. **`effective_profile_snapshot` defaults to `{}` instead of capturing the actual resolved profile** — spec requires a frozen JSON snapshot at plan creation time for audit/reproducibility; the code acknowledges this in a comment but doesn't fix it 2. **`agents plan list` is missing `--namespace/-n` filter** — spec requires `agents plan list [--namespace NS]`; the CLI command does not expose this option 3. **`VALID_PHASE_TRANSITIONS` allows `STRATEGIZE → EXECUTE` but not `STRATEGIZE → STRATEGIZE`** — the spec says Execute can revert to Strategize; the code correctly models this, but the `can_revert_to()` method on `Plan` uses `can_transition()` which doesn't allow `STRATEGIZE → STRATEGIZE` (needed for re-running Strategize after a revert from Execute) 4. **Invariant precedence: spec says plan > project > global (three-tier), but code models plan > action > project > global (four-tier)** — the spec glossary says "Action invariants are promoted to plan-level when the action is used, so the runtime precedence chain is three-tier: plan > project > global" — the code keeps action as a separate tier rather than promoting to plan level --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: ca-uat-tester

freemo referenced this issue

2026-04-03 17:23:39 +00:00

Proposal: improve ca-uat-tester — stop creating separate issues for every progress report #2369

freemo commented

2026-04-03 17:29:45 +00:00

Author

Owner

UAT Worker Starting — Skills System

Instance: uat-tester-skills-system
Focus Area: Skills System (skill definition, registration, invocation, composition, lifecycle)
Started: 2026-04-03

Testing Scope

SkillConfigSchema (YAML validation, env var interpolation, camelCase normalization)
Skill domain model (SkillResolver, cycle detection, include resolution)
SkillRegistry (registration, lookup, refresh, validation)
SkillService (add/remove/list/show/tools/refresh)
SkillContext (read-only enforcement, change tracking, resource resolution)
InlineToolExecutor (timeout, write guard, path validation)
AgentSkillLoader (progressive disclosure, SKILL.md parsing)
CLI commands (agents skill add/remove/list/show/tools/refresh)

Status

Currently running code-level analysis and runtime tests.

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: ca-uat-tester

## UAT Worker Starting — Skills System **Instance**: uat-tester-skills-system **Focus Area**: Skills System (skill definition, registration, invocation, composition, lifecycle) **Started**: 2026-04-03 ### Testing Scope - SkillConfigSchema (YAML validation, env var interpolation, camelCase normalization) - Skill domain model (SkillResolver, cycle detection, include resolution) - SkillRegistry (registration, lookup, refresh, validation) - SkillService (add/remove/list/show/tools/refresh) - SkillContext (read-only enforcement, change tracking, resource resolution) - InlineToolExecutor (timeout, write guard, path validation) - AgentSkillLoader (progressive disclosure, SKILL.md parsing) - CLI commands (agents skill add/remove/list/show/tools/refresh) ### Status Currently running code-level analysis and runtime tests. --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: ca-uat-tester

freemo commented

2026-04-03 17:31:59 +00:00

Author

Owner

UAT Worker Starting — CLI Commands Feature Area

Instance: uat-tester-cli-worker
Feature Area: CLI Commands
Started: 2026-04-03

Scope

Testing all CLI command groups against the specification:

Core commands: version, info, diagnostics, init
Command groups: project, plan, actor, action, resource, skill, tool, lsp, session, validation, invariant, automation-profile, config
Global options: --format, --data-dir, --config-path, -v verbosity
Shortcut commands: tell, build, apply

Method

Code-level analysis comparing implementation against specification
Runtime testing via uv run where feasible

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: ca-uat-tester

## UAT Worker Starting — CLI Commands Feature Area **Instance**: uat-tester-cli-worker **Feature Area**: CLI Commands **Started**: 2026-04-03 ### Scope Testing all CLI command groups against the specification: - Core commands: `version`, `info`, `diagnostics`, `init` - Command groups: `project`, `plan`, `actor`, `action`, `resource`, `skill`, `tool`, `lsp`, `session`, `validation`, `invariant`, `automation-profile`, `config` - Global options: `--format`, `--data-dir`, `--config-path`, `-v` verbosity - Shortcut commands: `tell`, `build`, `apply` ### Method - Code-level analysis comparing implementation against specification - Runtime testing via `uv run` where feasible --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: ca-uat-tester

freemo commented

2026-04-03 17:32:32 +00:00

Author

Owner

UAT Worker Complete — Plan Lifecycle Feature Area

Instance: uat-tester-plan-lifecycle
Focus Area: Plan Lifecycle (Action/Strategize/Execute/Apply phases)
Completed: 2026-04-03

Testing Summary

Code Analysis Coverage: 100% of plan lifecycle feature areas
Runtime Testing: N/A (code-level analysis only — environment constraints prevented runtime test execution)

Files Analyzed

File	Analysis
`src/cleveragents/domain/models/core/plan.py`	Full — PlanPhase, ProcessingState, VALID_PHASE_TRANSITIONS, Plan model, validators
`src/cleveragents/action/schema.py`	Full — ActionConfigSchema, argument validation, YAML loading
`src/cleveragents/domain/models/core/action.py`	Full — Action model, ActionState, ActionArgument
`src/cleveragents/application/services/plan_lifecycle_service.py`	Full — use_action, execute_plan, apply_plan, revert_plan, phase transitions
`src/cleveragents/cli/commands/plan.py`	Full — plan list, plan use, plan execute, plan apply, plan cancel
`docs/specification.md`	Plan lifecycle sections — phases, states, transitions, invariants, automation profiles
`features/plan_lifecycle_service.feature`	Full — BDD test coverage review
`features/plan_lifecycle_coverage.feature`	Full — coverage gap review

Features Tested

Feature	Status	Notes
Four-phase lifecycle (Action→Strategize→Execute→Apply)	✅ Correct	Phase ordering and transitions correctly implemented
Phase state machine (QUEUED→PROCESSING→COMPLETE/ERRORED)	✅ Correct	Per-phase state constraints enforced by validators
`plan use` creates plan in STRATEGIZE/QUEUED	✅ Correct	Matches spec
`plan execute` transitions STRATEGIZE/COMPLETE→EXECUTE/QUEUED	✅ Correct	Matches spec
`plan apply` transitions EXECUTE/COMPLETE→APPLY/QUEUED	✅ Correct	Matches spec
Apply terminal states (applied/constrained/errored/cancelled)	✅ Correct	Correctly modeled as ProcessingState in APPLY phase
Phase reversion (EXECUTE→STRATEGIZE, APPLY→STRATEGIZE)	✅ Correct	VALID_PHASE_TRANSITIONS correctly allows these
APPLIED/CANCELLED are permanently terminal	✅ Correct	`can_revert_to()` blocks these
MAX_REVERSIONS limit (3)	✅ Correct	Enforced in `can_revert_to()`
Action YAML schema validation	✅ Correct	Required fields, actor name format, argument types
Non-reusable action archived after use	✅ Correct	Implemented in `use_action()`
Automation profile precedence (plan>action>project>global)	✅ Correct	`_resolve_plan_profile_ref()` implements correctly
Invariant merging at plan creation	⚠️ Bug	Action invariants not promoted to plan-level (Issue #2392)
`effective_profile_snapshot` capture	⚠️ Bug	Always `{}` — not populated (Issue #2378)
`STRATEGIZE→STRATEGIZE` reversion	⚠️ Bug	Not in VALID_PHASE_TRANSITIONS (Issue #2380)
`agents plan list --namespace` filter	✅ Already filed	Issue already exists in tracker
Plan ULID identification	✅ Correct	ULID pattern validated
Subplan hierarchy (parent_plan_id, root_plan_id)	✅ Correct	Correctly modeled
Decision recording during phase transitions	✅ Correct	Silently skipped if no DecisionService
Event emission during phase transitions	✅ Correct	Silently skipped if no EventBus
Pre-flight guardrail checks before Strategize	✅ Correct	7 checks run before QUEUED→PROCESSING

Bugs Filed

Issue	Title	Severity
#2378	`effective_profile_snapshot` defaults to `{}` — audit trail incomplete	Medium
#2380	`Plan.can_revert_to()` rejects STRATEGIZE→STRATEGIZE — errored plans stuck	High
#2392	Action invariants stored as `InvariantSource.ACTION` instead of promoted to `PLAN`	Medium

No Issues Found (Confirmed Correct)

Phase ordering: Action → Strategize → Execute → Apply ✅
ProcessingState constraints per phase (APPLIED/CONSTRAINED only in APPLY, COMPLETE only in STRATEGIZE/EXECUTE) ✅
is_terminal property correctly identifies APPLIED, CANCELLED, CONSTRAINED, ERRORED ✅
can_transition_to_next_phase correctly requires COMPLETE state ✅
get_next_phase() returns correct next phase ✅
Action state lifecycle (available/archived, no draft) ✅
Automation profile resolution precedence chain ✅
Plan ULID generation and validation ✅
Project link alias uniqueness validation ✅
execution_env_priority requires execution_environment validation ✅

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: ca-uat-tester

## UAT Worker Complete — Plan Lifecycle Feature Area **Instance**: uat-tester-plan-lifecycle **Focus Area**: Plan Lifecycle (Action/Strategize/Execute/Apply phases) **Completed**: 2026-04-03 ### Testing Summary **Code Analysis Coverage**: 100% of plan lifecycle feature areas **Runtime Testing**: N/A (code-level analysis only — environment constraints prevented runtime test execution) ### Files Analyzed | File | Analysis | |------|----------| | `src/cleveragents/domain/models/core/plan.py` | Full — PlanPhase, ProcessingState, VALID_PHASE_TRANSITIONS, Plan model, validators | | `src/cleveragents/action/schema.py` | Full — ActionConfigSchema, argument validation, YAML loading | | `src/cleveragents/domain/models/core/action.py` | Full — Action model, ActionState, ActionArgument | | `src/cleveragents/application/services/plan_lifecycle_service.py` | Full — use_action, execute_plan, apply_plan, revert_plan, phase transitions | | `src/cleveragents/cli/commands/plan.py` | Full — plan list, plan use, plan execute, plan apply, plan cancel | | `docs/specification.md` | Plan lifecycle sections — phases, states, transitions, invariants, automation profiles | | `features/plan_lifecycle_service.feature` | Full — BDD test coverage review | | `features/plan_lifecycle_coverage.feature` | Full — coverage gap review | ### Features Tested | Feature | Status | Notes | |---------|--------|-------| | Four-phase lifecycle (Action→Strategize→Execute→Apply) | ✅ Correct | Phase ordering and transitions correctly implemented | | Phase state machine (QUEUED→PROCESSING→COMPLETE/ERRORED) | ✅ Correct | Per-phase state constraints enforced by validators | | `plan use` creates plan in STRATEGIZE/QUEUED | ✅ Correct | Matches spec | | `plan execute` transitions STRATEGIZE/COMPLETE→EXECUTE/QUEUED | ✅ Correct | Matches spec | | `plan apply` transitions EXECUTE/COMPLETE→APPLY/QUEUED | ✅ Correct | Matches spec | | Apply terminal states (applied/constrained/errored/cancelled) | ✅ Correct | Correctly modeled as ProcessingState in APPLY phase | | Phase reversion (EXECUTE→STRATEGIZE, APPLY→STRATEGIZE) | ✅ Correct | VALID_PHASE_TRANSITIONS correctly allows these | | APPLIED/CANCELLED are permanently terminal | ✅ Correct | `can_revert_to()` blocks these | | MAX_REVERSIONS limit (3) | ✅ Correct | Enforced in `can_revert_to()` | | Action YAML schema validation | ✅ Correct | Required fields, actor name format, argument types | | Non-reusable action archived after use | ✅ Correct | Implemented in `use_action()` | | Automation profile precedence (plan>action>project>global) | ✅ Correct | `_resolve_plan_profile_ref()` implements correctly | | Invariant merging at plan creation | ⚠️ Bug | Action invariants not promoted to plan-level (Issue #2392) | | `effective_profile_snapshot` capture | ⚠️ Bug | Always `{}` — not populated (Issue #2378) | | `STRATEGIZE→STRATEGIZE` reversion | ⚠️ Bug | Not in VALID_PHASE_TRANSITIONS (Issue #2380) | | `agents plan list --namespace` filter | ✅ Already filed | Issue already exists in tracker | | Plan ULID identification | ✅ Correct | ULID pattern validated | | Subplan hierarchy (parent_plan_id, root_plan_id) | ✅ Correct | Correctly modeled | | Decision recording during phase transitions | ✅ Correct | Silently skipped if no DecisionService | | Event emission during phase transitions | ✅ Correct | Silently skipped if no EventBus | | Pre-flight guardrail checks before Strategize | ✅ Correct | 7 checks run before QUEUED→PROCESSING | ### Bugs Filed | Issue | Title | Severity | |-------|-------|----------| | #2378 | `effective_profile_snapshot` defaults to `{}` — audit trail incomplete | Medium | | #2380 | `Plan.can_revert_to()` rejects STRATEGIZE→STRATEGIZE — errored plans stuck | High | | #2392 | Action invariants stored as `InvariantSource.ACTION` instead of promoted to `PLAN` | Medium | ### No Issues Found (Confirmed Correct) - Phase ordering: Action → Strategize → Execute → Apply ✅ - ProcessingState constraints per phase (APPLIED/CONSTRAINED only in APPLY, COMPLETE only in STRATEGIZE/EXECUTE) ✅ - `is_terminal` property correctly identifies APPLIED, CANCELLED, CONSTRAINED, ERRORED ✅ - `can_transition_to_next_phase` correctly requires COMPLETE state ✅ - `get_next_phase()` returns correct next phase ✅ - Action state lifecycle (available/archived, no draft) ✅ - Automation profile resolution precedence chain ✅ - Plan ULID generation and validation ✅ - Project link alias uniqueness validation ✅ - `execution_env_priority` requires `execution_environment` validation ✅ --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: ca-uat-tester

freemo commented

2026-04-03 17:33:13 +00:00

Author

Owner

UAT Worker Completed — API Endpoints Feature Area

Instance: uat-tester-api-endpoints
Feature Area: api-endpoints (REST API / A2A JSON-RPC endpoints)
Status: ✅ COMPLETE

Summary

Performed comprehensive code-level analysis of the A2A protocol layer against the specification. Tested:

A2aLocalFacade dispatch routing and handler coverage
A2aRequest / A2aResponse wire format compliance
ASGI health endpoint implementation
A2aEventQueue and SseEventFormatter behavior
A2aVersionNegotiator version negotiation
ServerConnectionConfig validation
A2aHttpTransport and stub client implementations

Features Tested: 8/8

Feature	Status	Notes
JSON-RPC 2.0 wire format (A2aRequest/A2aResponse)	✅ PASS	Correct field names, validators, mutual exclusion
ASGI health/liveness/readiness endpoints	✅ PASS	`/live`, `/ready`, `/health`, `/` all correct
A2A facade plan lifecycle operations	⚠️ PARTIAL	Core ops work; `plan/cancel` drops reason, `plan/rollback` ignores checkpoint_id
A2A facade registry operations	❌ FAIL	Only `list` implemented; `show`, `add`, `update`, `remove` missing for all entity types
A2A facade session operations	❌ FAIL	Only `session.create`/`session.close` exist; `list`, `show`, `delete`, `export`, `import`, `tell` missing
A2A event queue and SSE formatting	✅ PASS	Local pub/sub works; SSE format correct
A2A version negotiation	✅ PASS	Version 1.0 negotiation works
Server connection config validation	✅ PASS	URL and namespace validation correct

Bugs Filed: 4 new issues

Issue	Title	Severity
#2384	`A2aLocalFacade` missing registry `show`, `add`, `update`, `remove` extension methods	High
#2385	`_cleveragents/plan/cancel` ignores `reason` parameter	Medium
#2387	`_cleveragents/plan/rollback` ignores `checkpoint_id` parameter	High
#2396	`A2aLocalFacade` missing session management A2A operations	High

Previously Reported (not re-filed)

The following related issues were already open and were not duplicated:

ASGI app missing A2A JSON-RPC endpoint and SSE streaming endpoint
A2aHttpTransport raises A2aNotAvailableError on all operations
_cleveragents/namespace/* methods are unimplemented stubs
_cleveragents/sync/* methods return stub not_implemented
Local mode A2A stdio transport not implemented

Runtime Test Coverage

Code analysis: 100% of identified features
Runtime tests: 0% (bash execution restricted in this environment)

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: ca-uat-tester

## UAT Worker Completed — API Endpoints Feature Area **Instance**: uat-tester-api-endpoints **Feature Area**: api-endpoints (REST API / A2A JSON-RPC endpoints) **Status**: ✅ COMPLETE --- ### Summary Performed comprehensive code-level analysis of the A2A protocol layer against the specification. Tested: - `A2aLocalFacade` dispatch routing and handler coverage - `A2aRequest` / `A2aResponse` wire format compliance - ASGI health endpoint implementation - `A2aEventQueue` and `SseEventFormatter` behavior - `A2aVersionNegotiator` version negotiation - `ServerConnectionConfig` validation - `A2aHttpTransport` and stub client implementations ### Features Tested: 8/8 | Feature | Status | Notes | |---|---|---| | JSON-RPC 2.0 wire format (A2aRequest/A2aResponse) | ✅ PASS | Correct field names, validators, mutual exclusion | | ASGI health/liveness/readiness endpoints | ✅ PASS | `/live`, `/ready`, `/health`, `/` all correct | | A2A facade plan lifecycle operations | ⚠️ PARTIAL | Core ops work; `plan/cancel` drops reason, `plan/rollback` ignores checkpoint_id | | A2A facade registry operations | ❌ FAIL | Only `list` implemented; `show`, `add`, `update`, `remove` missing for all entity types | | A2A facade session operations | ❌ FAIL | Only `session.create`/`session.close` exist; `list`, `show`, `delete`, `export`, `import`, `tell` missing | | A2A event queue and SSE formatting | ✅ PASS | Local pub/sub works; SSE format correct | | A2A version negotiation | ✅ PASS | Version 1.0 negotiation works | | Server connection config validation | ✅ PASS | URL and namespace validation correct | ### Bugs Filed: 4 new issues | Issue | Title | Severity | |---|---|---| | #2384 | `A2aLocalFacade` missing registry `show`, `add`, `update`, `remove` extension methods | High | | #2385 | `_cleveragents/plan/cancel` ignores `reason` parameter | Medium | | #2387 | `_cleveragents/plan/rollback` ignores `checkpoint_id` parameter | High | | #2396 | `A2aLocalFacade` missing session management A2A operations | High | ### Previously Reported (not re-filed) The following related issues were already open and were not duplicated: - ASGI app missing A2A JSON-RPC endpoint and SSE streaming endpoint - `A2aHttpTransport` raises `A2aNotAvailableError` on all operations - `_cleveragents/namespace/*` methods are unimplemented stubs - `_cleveragents/sync/*` methods return stub `not_implemented` - Local mode A2A stdio transport not implemented ### Runtime Test Coverage - Code analysis: 100% of identified features - Runtime tests: 0% (bash execution restricted in this environment) --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: ca-uat-tester

freemo commented

2026-04-03 17:34:10 +00:00

Author

Owner

UAT Worker Starting — Sandbox and Checkpoint Model

Instance: uat-sandbox-worker (pool: uat-pool-1)
Feature Area: sandbox-and-checkpoint-model
Working Directory: /app (read-only analysis mode — bash execution restricted)
Start Time: 2026-04-03

Scope

Testing the following spec-required behaviors:

Safety sandbox model (Execute phase isolation)
Checkpoint creation and restoration
Reversible execution
Safety profiles (require_sandbox, require_checkpoints, unsafe_tools gating)
agents plan rollback <PLAN_ID> <CHECKPOINT_ID> command
Apply phase (sandbox → real resources promotion)

Will perform code-level analysis against the specification. Filing bugs for any deviations found.

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: ca-uat-tester

## UAT Worker Starting — Sandbox and Checkpoint Model **Instance**: uat-sandbox-worker (pool: uat-pool-1) **Feature Area**: sandbox-and-checkpoint-model **Working Directory**: /app (read-only analysis mode — bash execution restricted) **Start Time**: 2026-04-03 ### Scope Testing the following spec-required behaviors: - Safety sandbox model (Execute phase isolation) - Checkpoint creation and restoration - Reversible execution - Safety profiles (require_sandbox, require_checkpoints, unsafe_tools gating) - `agents plan rollback <PLAN_ID> <CHECKPOINT_ID>` command - Apply phase (sandbox → real resources promotion) Will perform code-level analysis against the specification. Filing bugs for any deviations found. --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: ca-uat-tester

freemo referenced this issue

2026-04-03 17:34:44 +00:00

[CA-GROOMER] Backlog Groomer — Session State (groomer-1) #2386

freemo commented

2026-04-03 17:35:12 +00:00

Author

Owner

UAT Pool Supervisor — Progress Report (Cycle 12)

Time: 2026-04-03 UTC (~12 minutes elapsed)
HEAD: 8866c58bd4

Worker Status

Active: 10/10
Tested areas: 2/12 (plan-lifecycle ✅, api-endpoints ✅)
Coverage: 17%
Bugs filed so far: ~26 (across all workers)

Completed Workers

Feature Area	Bugs Filed	Key Issues
plan-lifecycle	3	#2378, #2380, #2392
api-endpoints	4	#2384, #2385, #2387, #2396

Active Workers (10 running)

Session	Feature Area
ses_2aba47eaeffeBGrz4sDDwFTFi0	projects-and-resources
ses_2aba47b28ffeGuDOO3QPeHtf8m	actor-system
ses_2aba475ceffebf0G2JGXmamQ8K	tool-sources-and-security
ses_2aba47195ffeXBwLPvRw587hbG	skills-system
ses_2aba46eccffeJWZHDBOONksK2y	context-management
ses_2aba467bbffe46WVyrQJMUacmQ	cli-commands
ses_2aba463fcffexweqFuQKZpwIDD	data-models-and-schemas
ses_2aba460b8ffeuGQUxrwRu1lwgQ	configuration-and-environment
ses_2ab98326effeLQHi65FmwdULIh	sandbox-and-checkpoint-model
ses_2ab96cd77ffefiJZ859494WxCn	testing-infrastructure

Notable Bugs Found (sample)

#2395: 31 # type: ignore[misc] comments violating no-type-ignore rule
#2393: agents actor add silently overwrites existing actors
#2392: Action invariants stored incorrectly (ACTION vs PLAN source)
#2391: Project invariants silently lost on read-back
#2390: ContextFragment.detail_depth constraint too restrictive
#2389: ContextTierService doesn't enforce warm/cold tier budgets
#2388: agents actor show missing spec-required fields
#2387: Plan rollback ignores checkpoint_id parameter
#2385: Plan cancel ignores reason parameter
#2384: A2A facade missing CRUD operations for 11 entity types
#2383: Resource CLI bypasses service layer
#2382: ACMS pipeline doesn't forward full ContextRequest
#2380: Plan.can_revert_to() rejects valid STRATEGIZE→STRATEGIZE reversion
#2379: Non-existent default automation profile name
#2378: effective_profile_snapshot always empty dict

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: ca-uat-tester

## UAT Pool Supervisor — Progress Report (Cycle 12) **Time**: 2026-04-03 UTC (~12 minutes elapsed) **HEAD**: 8866c58bd49d ### Worker Status - Active: 10/10 - Tested areas: 2/12 (plan-lifecycle ✅, api-endpoints ✅) - Coverage: 17% - Bugs filed so far: ~26 (across all workers) ### Completed Workers | Feature Area | Bugs Filed | Key Issues | |---|---|---| | plan-lifecycle | 3 | #2378, #2380, #2392 | | api-endpoints | 4 | #2384, #2385, #2387, #2396 | ### Active Workers (10 running) | Session | Feature Area | |---|---| | ses_2aba47eaeffeBGrz4sDDwFTFi0 | projects-and-resources | | ses_2aba47b28ffeGuDOO3QPeHtf8m | actor-system | | ses_2aba475ceffebf0G2JGXmamQ8K | tool-sources-and-security | | ses_2aba47195ffeXBwLPvRw587hbG | skills-system | | ses_2aba46eccffeJWZHDBOONksK2y | context-management | | ses_2aba467bbffe46WVyrQJMUacmQ | cli-commands | | ses_2aba463fcffexweqFuQKZpwIDD | data-models-and-schemas | | ses_2aba460b8ffeuGQUxrwRu1lwgQ | configuration-and-environment | | ses_2ab98326effeLQHi65FmwdULIh | sandbox-and-checkpoint-model | | ses_2ab96cd77ffefiJZ859494WxCn | testing-infrastructure | ### Notable Bugs Found (sample) - #2395: 31 `# type: ignore[misc]` comments violating no-type-ignore rule - #2393: `agents actor add` silently overwrites existing actors - #2392: Action invariants stored incorrectly (ACTION vs PLAN source) - #2391: Project invariants silently lost on read-back - #2390: `ContextFragment.detail_depth` constraint too restrictive - #2389: `ContextTierService` doesn't enforce warm/cold tier budgets - #2388: `agents actor show` missing spec-required fields - #2387: Plan rollback ignores `checkpoint_id` parameter - #2385: Plan cancel ignores `reason` parameter - #2384: A2A facade missing CRUD operations for 11 entity types - #2383: Resource CLI bypasses service layer - #2382: ACMS pipeline doesn't forward full ContextRequest - #2380: `Plan.can_revert_to()` rejects valid STRATEGIZE→STRATEGIZE reversion - #2379: Non-existent default automation profile name - #2378: `effective_profile_snapshot` always empty dict --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: ca-uat-tester

freemo commented

2026-04-03 17:35:14 +00:00

Author

Owner

UAT Worker Complete — Context Management (ACMS/UKO)

Instance: uat-context-mgmt-worker
Feature Area: Context Management (ACMS + UKO)
Completed: 2026-04-03

Summary

Code-level analysis of the ACMS/UKO context management system against the specification. All major sub-areas were reviewed.

Areas Tested

Sub-Area	Status	Notes
UKO Vocabulary (Layers 0-3)	✅ Implemented	Layers 0-3 present with correct namespace IRIs and class inventories
UKO Vocabulary Registry	✅ Implemented	`VocabularyRegistry` with prefix/IRI lookup, sorted list
UKO Detail Level Maps	✅ Implemented	`DetailLevelMap` with inheritance, cycle detection, `MappingProxyType`
UKO Indexer	✅ Implemented	`UKOIndexer` with graph/text/vector backends, provenance, idempotency
UKO Query Interface	✅ Implemented	`UKOQueryInterface` with layer inference, implicit relations
ACMS Pipeline (10-stage)	⚠️ Partial	Pipeline exists but CRP fields not forwarded to strategies (Bug #2382)
Context Request Protocol (CRP)	⚠️ Partial	`ContextRequest` model complete; not used by strategies (Bug #2382)
Context Strategies	⚠️ Partial	6 built-in strategies use wrong protocol (Bug #2375)
Tiered Storage (hot/warm/cold)	⚠️ Partial	Hot tier enforced; warm/cold limits not enforced (Bug #2389)
Skeleton Compressor	✅ Implemented	`DepthReductionCompressor` with domain-aware depth reduction
CLI `project context set/show/inspect/simulate`	✅ Implemented	Full ACMS config management via CLI
CLI `actor context add/remove/list/show`	✅ Implemented	File-based context management
Temporal Versioning (UKO)	✅ Implemented	`TemporalMetadata`, `RevisionChain`, `TemporalBackend` protocol
Per-Actor Context Views	✅ Implemented	`ActorContextView` with role-based tier visibility

Bugs Filed (5 total)

#	Issue	Priority	Description
1	#2375	High	Dual incompatible `ContextStrategy` protocols — domain model protocol never used
2	#2382	High	`ACMSPipeline.assemble()` passes only `{"strategy": name}` dict — full `ContextRequest` never forwarded to strategies
3	#2390	High	`ContextFragment.detail_depth` hardcoded `le=9` — prevents `uko-doc:` (max 10) and `uko-data:` (max 11) fragments
4	#2389	Medium	`ContextTierService` warm/cold tier capacity limits not enforced — tiers grow unboundedly
5	#2402	Medium	`DefaultDepthResolver` is a no-op — named depth levels from `ContextRequest.depth` never resolved

Coverage

Code Analysis Coverage: 100% of ACMS/UKO sub-areas reviewed
Runtime Test Coverage: 0% (environment setup not performed — code analysis only)
Critical: 0 | High: 3 | Medium: 2 | Low: 0

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: ca-uat-tester

## UAT Worker Complete — Context Management (ACMS/UKO) **Instance**: uat-context-mgmt-worker **Feature Area**: Context Management (ACMS + UKO) **Completed**: 2026-04-03 --- ### Summary Code-level analysis of the ACMS/UKO context management system against the specification. All major sub-areas were reviewed. ### Areas Tested | Sub-Area | Status | Notes | |---|---|---| | UKO Vocabulary (Layers 0-3) | ✅ Implemented | Layers 0-3 present with correct namespace IRIs and class inventories | | UKO Vocabulary Registry | ✅ Implemented | `VocabularyRegistry` with prefix/IRI lookup, sorted list | | UKO Detail Level Maps | ✅ Implemented | `DetailLevelMap` with inheritance, cycle detection, `MappingProxyType` | | UKO Indexer | ✅ Implemented | `UKOIndexer` with graph/text/vector backends, provenance, idempotency | | UKO Query Interface | ✅ Implemented | `UKOQueryInterface` with layer inference, implicit relations | | ACMS Pipeline (10-stage) | ⚠️ Partial | Pipeline exists but CRP fields not forwarded to strategies (Bug #2382) | | Context Request Protocol (CRP) | ⚠️ Partial | `ContextRequest` model complete; not used by strategies (Bug #2382) | | Context Strategies | ⚠️ Partial | 6 built-in strategies use wrong protocol (Bug #2375) | | Tiered Storage (hot/warm/cold) | ⚠️ Partial | Hot tier enforced; warm/cold limits not enforced (Bug #2389) | | Skeleton Compressor | ✅ Implemented | `DepthReductionCompressor` with domain-aware depth reduction | | CLI `project context set/show/inspect/simulate` | ✅ Implemented | Full ACMS config management via CLI | | CLI `actor context add/remove/list/show` | ✅ Implemented | File-based context management | | Temporal Versioning (UKO) | ✅ Implemented | `TemporalMetadata`, `RevisionChain`, `TemporalBackend` protocol | | Per-Actor Context Views | ✅ Implemented | `ActorContextView` with role-based tier visibility | ### Bugs Filed (5 total) | # | Issue | Priority | Description | |---|---|---|---| | 1 | [#2375](https://git.cleverthis.com/cleveragents/cleveragents-core/issues/2375) | High | Dual incompatible `ContextStrategy` protocols — domain model protocol never used | | 2 | [#2382](https://git.cleverthis.com/cleveragents/cleveragents-core/issues/2382) | High | `ACMSPipeline.assemble()` passes only `{"strategy": name}` dict — full `ContextRequest` never forwarded to strategies | | 3 | [#2390](https://git.cleverthis.com/cleveragents/cleveragents-core/issues/2390) | High | `ContextFragment.detail_depth` hardcoded `le=9` — prevents `uko-doc:` (max 10) and `uko-data:` (max 11) fragments | | 4 | [#2389](https://git.cleverthis.com/cleveragents/cleveragents-core/issues/2389) | Medium | `ContextTierService` warm/cold tier capacity limits not enforced — tiers grow unboundedly | | 5 | [#2402](https://git.cleverthis.com/cleveragents/cleveragents-core/issues/2402) | Medium | `DefaultDepthResolver` is a no-op — named depth levels from `ContextRequest.depth` never resolved | ### Coverage - **Code Analysis Coverage**: 100% of ACMS/UKO sub-areas reviewed - **Runtime Test Coverage**: 0% (environment setup not performed — code analysis only) - **Critical**: 0 | **High**: 3 | **Medium**: 2 | **Low**: 0 --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: ca-uat-tester

freemo commented

2026-04-03 17:35:34 +00:00

Author

Owner

UAT Worker Starting — Testing Infrastructure

Instance: uat-tester-testing-infrastructure
Feature Area: Testing Infrastructure
Focus: Behave BDD framework, Robot Framework integration tests, nox sessions, coverage reporting, test organization

Starting analysis of testing infrastructure against spec requirements.

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: ca-uat-tester

## UAT Worker Starting — Testing Infrastructure **Instance**: uat-tester-testing-infrastructure **Feature Area**: Testing Infrastructure **Focus**: Behave BDD framework, Robot Framework integration tests, nox sessions, coverage reporting, test organization Starting analysis of testing infrastructure against spec requirements. --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: ca-uat-tester

freemo referenced this issue

2026-04-03 17:36:29 +00:00

UAT: Global CLI options --data-dir and --config-path are not implemented — spec requires both #2409

freemo commented

2026-04-03 17:36:53 +00:00

Author

Owner

UAT Worker Report — Projects & Resources Feature Area

Instance: uat-tester-projects-and-resources
Feature Area: projects-and-resources
Completed: 2026-04-03

Summary

Completed UAT testing of the Projects and Resources feature area. Performed deep code analysis of:

src/cleveragents/cli/commands/project.py — Project CLI commands
src/cleveragents/cli/commands/resource.py — Resource CLI commands
src/cleveragents/domain/models/core/project.py — NamespacedProject domain model
src/cleveragents/domain/models/core/resource.py — Resource domain model
src/cleveragents/domain/models/core/resource_type.py — ResourceTypeSpec domain model
src/cleveragents/application/services/resource_registry_service.py — Resource registry service
src/cleveragents/application/services/_resource_registry_ops.py — Resource instance operations
src/cleveragents/application/services/_resource_registry_dag.py — DAG operations
src/cleveragents/infrastructure/database/models.py — NamespacedProjectModel ORM model
src/cleveragents/infrastructure/database/repositories.py — NamespacedProjectRepository
Feature test files: project_cli.feature, resource_cli.feature, resource_dag.feature, project_service.feature, resource_type_builtins.feature, resource_handler_crud.feature

Features Tested

Feature	Code Analysis	Status
`agents project create`	✅	Functional with bug (invariants not persisted through domain model)
`agents project list`	✅	Functional
`agents project show`	✅	Functional (but missing invariants display)
`agents project delete`	✅	Functional
`agents project link-resource`	✅	Functional
`agents project unlink-resource`	✅	Functional
`agents resource add`	✅	Functional (--update bypasses service layer)
`agents resource list`	✅	Functional
`agents resource show`	✅	Functional
`agents resource remove`	✅	Bug — bypasses service layer, blocks on edges
`agents resource tree`	✅	Functional
`agents resource inspect`	✅	Functional
`agents resource link-child`	✅	Functional
`agents resource unlink-child`	✅	Functional
`agents resource type add`	✅	Functional (--update not implemented)
`agents resource type list`	✅	Functional
`agents resource type show`	✅	Functional
`agents resource type remove`	✅	Functional
Resource DAG (cycle detection)	✅	Functional
Resource type inheritance (ADR-042)	✅	Functional
Built-in resource types bootstrap	✅	Functional
Namespaced project model validation	✅	Functional
Project namespace validation	✅	Functional

Bugs Filed (3)

Issue	Title	Priority
#2383	`agents resource remove` and `agents resource add --update` bypass `ResourceRegistryService` — CLI directly manipulates infrastructure DB models	High
#2391	`agents project create --invariant` stores invariants via raw SQL but `NamespacedProject` domain model has no `invariants` field — invariants are silently lost on read-back	High
#2401	`agents resource type add --update` flag is documented but always aborts with "Update mode is not yet fully supported"	Medium
#2407	`agents resource remove` refuses to remove resources with DAG edges — no `--force` option provided	Medium

Key Findings

Critical architectural violation (#2383): The agents resource remove command and --update flag in agents resource add bypass the ResourceRegistryService entirely, directly importing and manipulating ResourceModel and ResourceEdgeModel from the infrastructure layer. The service has no remove_resource() method.
Data loss bug (#2391): Project invariants specified via --invariant are stored via raw SQL but the NamespacedProject domain model has no invariants field. The NamespacedProjectModel.to_domain() method silently drops invariants when converting from DB to domain model. from_domain() always sets invariants_json=json.dumps([]).
Unimplemented feature (#2401): The --update flag on agents resource type add is shown in CLI help but explicitly not implemented — the code prints "Update mode is not yet fully supported" and aborts.
Missing --force option (#2407): agents resource remove hard-blocks removal of any resource with DAG edges, with no --force option to cascade-delete edges.

What Was NOT Reported (Already Covered or Not Bugs)

agents resource stop rejecting container-instance type — already reported as existing UAT bug
Devcontainer auto-discovery not wired — already reported as existing UAT bug
resource add storing path in both location and properties["path"] — minor redundancy, not a functional bug
Resource name format validation in register_resource — the spec doesn't strictly require namespaced format for resource names

Automated by CleverAgents Bot
Supervisor: UAT Testing | Agent: ca-uat-tester

## UAT Worker Report — Projects & Resources Feature Area **Instance**: uat-tester-projects-and-resources **Feature Area**: projects-and-resources **Completed**: 2026-04-03 --- ### Summary Completed UAT testing of the **Projects and Resources** feature area. Performed deep code analysis of: - `src/cleveragents/cli/commands/project.py` — Project CLI commands - `src/cleveragents/cli/commands/resource.py` — Resource CLI commands - `src/cleveragents/domain/models/core/project.py` — `NamespacedProject` domain model - `src/cleveragents/domain/models/core/resource.py` — `Resource` domain model - `src/cleveragents/domain/models/core/resource_type.py` — `ResourceTypeSpec` domain model - `src/cleveragents/application/services/resource_registry_service.py` — Resource registry service - `src/cleveragents/application/services/_resource_registry_ops.py` — Resource instance operations - `src/cleveragents/application/services/_resource_registry_dag.py` — DAG operations - `src/cleveragents/infrastructure/database/models.py` — `NamespacedProjectModel` ORM model - `src/cleveragents/infrastructure/database/repositories.py` — `NamespacedProjectRepository` - Feature test files: `project_cli.feature`, `resource_cli.feature`, `resource_dag.feature`, `project_service.feature`, `resource_type_builtins.feature`, `resource_handler_crud.feature` --- ### Features Tested | Feature | Code Analysis | Status | |---|---|---| | `agents project create` | ✅ | Functional with bug (invariants not persisted through domain model) | | `agents project list` | ✅ | Functional | | `agents project show` | ✅ | Functional (but missing invariants display) | | `agents project delete` | ✅ | Functional | | `agents project link-resource` | ✅ | Functional | | `agents project unlink-resource` | ✅ | Functional | | `agents resource add` | ✅ | Functional (--update bypasses service layer) | | `agents resource list` | ✅ | Functional | | `agents resource show` | ✅ | Functional | | `agents resource remove` | ✅ | **Bug** — bypasses service layer, blocks on edges | | `agents resource tree` | ✅ | Functional | | `agents resource inspect` | ✅ | Functional | | `agents resource link-child` | ✅ | Functional | | `agents resource unlink-child` | ✅ | Functional | | `agents resource type add` | ✅ | Functional (--update not implemented) | | `agents resource type list` | ✅ | Functional | | `agents resource type show` | ✅ | Functional | | `agents resource type remove` | ✅ | Functional | | Resource DAG (cycle detection) | ✅ | Functional | | Resource type inheritance (ADR-042) | ✅ | Functional | | Built-in resource types bootstrap | ✅ | Functional | | Namespaced project model validation | ✅ | Functional | | Project namespace validation | ✅ | Functional | --- ### Bugs Filed (3) | Issue | Title | Priority | |---|---|---| | #2383 | `agents resource remove` and `agents resource add --update` bypass `ResourceRegistryService` — CLI directly manipulates infrastructure DB models | High | | #2391 | `agents project create --invariant` stores invariants via raw SQL but `NamespacedProject` domain model has no `invariants` field — invariants are silently lost on read-back | High | | #2401 | `agents resource type add --update` flag is documented but always aborts with "Update mode is not yet fully supported" | Medium | | #2407 | `agents resource remove` refuses to remove resources with DAG edges — no `--force` option provided | Medium | --- ### Key Findings 1. **Critical architectural violation** (#2383): The `agents resource remove` command and `--update` flag in `agents resource add` bypass the `ResourceRegistryService` entirely, directly importing and manipulating `ResourceModel` and `ResourceEdgeModel` from the infrastructure layer. The service has no `remove_resource()` method. 2. **Data loss bug** (#2391): Project invariants specified via `--invariant` are stored via raw SQL but the `NamespacedProject` domain model has no `invariants` field. The `NamespacedProjectModel.to_domain()` method silently drops invariants when converting from DB to domain model. `from_domain()` always sets `invariants_json=json.dumps([])`. 3. **Unimplemented feature** (#2401): The `--update` flag on `agents resource type add` is shown in CLI help but explicitly not implemented — the code prints "Update mode is not yet fully supported" and aborts. 4. **Missing `--force` option** (#2407): `agents resource remove` hard-blocks removal of any resource with DAG edges, with no `--force` option to cascade-delete edges. --- ### What Was NOT Reported (Already Covered or Not Bugs) - `agents resource stop` rejecting `container-instance` type — already reported as existing UAT bug - Devcontainer auto-discovery not wired — already reported as existing UAT bug - `resource add` storing `path` in both `location` and `properties["path"]` — minor redundancy, not a functional bug - Resource name format validation in `register_resource` — the spec doesn't strictly require namespaced format for resource names --- **Automated by CleverAgents Bot** Supervisor: UAT Testing | Agent: ca-uat-tester

freemo referenced this issue

2026-04-03 17:39:45 +00:00

UAT: Global -v verbosity flag (repeatable up to 5x) is not implemented as a global option — spec requires -v... on all commands #2415