test(integration): workflow example 7 — CI/CD integration, automated PR review and fix (ci profile)

Implemented Robot Framework integration test suite validating the CI/CD
workflow (Specification Example 7). Tests cover ci-profile configuration
(automation-profile, format, log level), idempotent resource and project
registration with duplicate-detection assertions, three validation tools
(ci-lint, ci-typecheck, ci-tests) registration and resource attachment
via ToolRegistryService, action creation with typed arguments and
invariants per spec Step 2, plan lifecycle with phase-by-phase completion
through all phases (strategize, execute, apply) until terminal applied
state, and JSON output structure verification including plan_id, phase,
state, action, projects, and arguments fields.

Post-review fixes applied (round 1):
- H1: Fix truncated invariant #2 text to match spec line 39051
- H2: Fix definition_of_done to match spec's 5-bullet-point format
- M1: Register all 3 spec validations (ci-lint, ci-typecheck, ci-tests)
- M3: Add branch property to resource registration per spec Step 3
- M4: Replace phantom resource_id with real registered resource
- M5: Add projects and arguments field assertions to JSON output test
- M7: Replace 'polling-based' with 'phase-by-phase' in docs/comments
- L1: Use timezone-aware datetime (timezone.utc)
- L2: Use tempfile for resource path instead of hardcoded /tmp
- L3: Clarify json_output() docstring to reflect as_cli_dict() scope
- L4: Tighten attachment count assertion from >= 1 to == 1 (now == 3)

Post-review fixes applied (round 2):
- M2: Use Setup Test Environment With Database Isolation for pabot safety
- L1: Replace remaining hardcoded /tmp paths with tempfile (validation_attach,
  resource_idempotent duplicate registration)
- L2: Fix stale 'Polling-based' comment missed in round 1
- L5: Align action description with spec line 39027
- M1/L3/L6: Document automation_profile propagation gap in use_action()
  with TODO for when production code wires the field onto Plan
- L4: Add comment explaining local actor name deviation from spec

Includes CHANGELOG update describing the integration test scope.

ISSUES CLOSED: #771

benchmarks/decision_correction_model_bench.py

@@ -106,7 +106,7 @@ class CorrectionServiceSuite:
         """Benchmark creating a correction request."""
         self.service.request_correction(
             plan_id="plan-bench",
-            decision_id="DEC-001",
+            target_decision_id="DEC-001",
             mode=CorrectionMode.REVERT,
             guidance="Benchmark correction",
         )
@@ -115,7 +115,7 @@ class CorrectionServiceSuite:
         """Benchmark impact analysis (stub)."""
         req = self.service.request_correction(
             plan_id="plan-bench",
-            decision_id="DEC-001",
+            target_decision_id="DEC-001",
             mode=CorrectionMode.REVERT,
             guidance="Benchmark analysis",
         )
@@ -125,7 +125,7 @@ class CorrectionServiceSuite:
         """Benchmark correction execution (stub)."""
         req = self.service.request_correction(
             plan_id="plan-bench",
-            decision_id="DEC-001",
+            target_decision_id="DEC-001",
             mode=CorrectionMode.REVERT,
             guidance="Benchmark execution",
         )

benchmarks/decision_di_bench.py

@@ -44,7 +44,7 @@ except ModuleNotFoundError:
     from cleveragents.infrastructure.database.unit_of_work import UnitOfWork
 
 
-_PLAN_ID = "01HV00000000000000DIBENCH1"
+_PLAN_ID = "01HV00000000000000DJBENCH1"
 
 
 def _make_uow() -> UnitOfWork:

benchmarks/phase_reversion_bench.py

@@ -119,16 +119,16 @@ class AutoRevertSuite:
         )
         plan_id = plan.identity.plan_id
         self.service.start_strategize(plan_id)
+        # complete_strategize auto-progresses to EXECUTE with "ci" profile
         self.service.complete_strategize(plan_id)
         plan = self.service.get_plan(plan_id)
         if plan.phase != PlanPhase.EXECUTE:
             self.service.execute_plan(plan_id)
         self.service.start_execute(plan_id)
+        # complete_execute auto-progresses to APPLY with "ci" profile
         self.service.complete_execute(plan_id)
-        self.service.apply_plan(plan_id)
         self.service.start_apply(plan_id)
-        plan = self.service.get_plan(plan_id)
-        plan.processing_state = ProcessingState.CONSTRAINED
+        self.service.constrain_apply(plan_id, "benchmark constraint")
         self.service.try_auto_revert_from_apply(plan_id, "benchmark")
 
 

benchmarks/resource_cli_tree_bench.py

@@ -75,10 +75,10 @@ class ResourceTreeSuite:
         """Set up mock service."""
         root = _mock_resource()
         child1 = _mock_resource(
-            "01HBENCH0000000000CHILD001", "local/child-1", "fs-directory"
+            "01HBENCH0000000000CHJKD001", "local/child-1", "fs-directory"
         )
         child2 = _mock_resource(
-            "01HBENCH0000000000CHILD002", "local/child-2", "fs-directory"
+            "01HBENCH0000000000CHJKD002", "local/child-2", "fs-directory"
         )
         tree = [
             _mock_tree_node(

benchmarks/session_list_bench.py

@@ -61,6 +61,11 @@ class SessionListDISuite:
             bind=self._engine,
             expire_on_commit=False,
         )
+        # Pre-initialise attributes used by per-method setup hooks so that
+        # benchmarks still work even when the ASV runner skips per-method
+        # setup_<name>() callbacks (e.g. in fork-server mode).
+        self._empty_svc = self._make_service()
+        self._fresh_engine()
 
     def teardown(self) -> None:
         self._engine.dispose()

benchmarks/unified_context_models_bench.py

@@ -84,7 +84,7 @@ class TimeUnifiedModelCreation:
     def time_context_payload(self) -> None:
         """Create core ContextPayload (inherits CRP AssembledContext)."""
         for _ in range(1000):
-            ContextPayload(plan_id="01JQBENCHPN00000000000000AA")
+            ContextPayload(plan_id="01JQBENCHPN0000000000000AA")
 
 
 class TimeIsinstanceChecks:
@@ -104,7 +104,7 @@ class TimeIsinstanceChecks:
         )
         self.budget = ContextBudget(max_tokens=4096, reserved_tokens=512)
         self.payload = ContextPayload(
-            plan_id="01JQBENCHPN00000000000000AA",
+            plan_id="01JQBENCHPN0000000000000AA",
         )
 
     def time_isinstance_provenance(self) -> None: