5 Commits

Author	SHA1	Message	Date
freemo	7fb3fc76c8	fix(plan-lifecycle): add rollback_plan method to PlanLifecycleService ... - What was implemented - Added PLAN_ROLLED_BACK event type to the EventType enum at src/cleveragents/infrastructure/events/types.py to properly represent successful rollbacks in the domain model. - Implemented rollback_plan(plan_id: str, checkpoint_id: str) -> RollbackResult in PlanLifecycleService (src/cleveragents/application/services/plan_lifecycle_service.py) with: - Plan state validation: rejects rollback when the plan is in terminal APPLIED or CANCELLED states. - Delegation to CheckpointService.selective_rollback() to perform the actual rollback logic and obtain a RollbackResult. - Emission of PLAN_ROLLED_BACK as a domain event to reflect the completed rollback. - checkpoint_service is accepted as an optional constructor parameter; if not provided, a PlanError is raised to preserve backward compatibility. - Updated CLI behavior in src/cleveragents/cli/commands/plan.py so agents plan rollback routes through PlanLifecycleService.rollback_plan() rather than calling CheckpointService.selective_rollback() directly. - Updated PlanLifecycleService module docstring to include rollback_plan in the documented API. - Added Behave feature file features/plan_lifecycle_rollback.feature with 11 scenarios covering state validation, domain events, and delegation. - Added step implementations in features/steps/plan_lifecycle_rollback_steps.py to support the new scenarios. - Key design decisions - rollback_plan returns RollbackResult (the same result type produced by CheckpointService.selective_rollback) so the CLI can display rollback details consistently. - Terminal states APPLIED and CANCELLED are disallowed for rollback to prevent inconsistent or invalid state transitions. - checkpoint_service is optional in the PlanLifecycleService constructor; when omitted (None), a PlanError is raised to retain backward compatibility while signaling explicit dependency requirements. - CLI UI remains powered by CheckpointService for metadata enrichment (e.g., confirmation prompts), but the actual rollback action is performed via PlanLifecycleService to ensure proper domain workflow and event emission. - Technical implications - All rollback logic now flows through the domain service layer (PlanLifecycleService) to preserve invariants and emit domain events, rather than allowing ad-hoc UI routes to bypass service validation. - The UI can still retrieve checkpoint metadata for user confirmation, but the operation that modifies state uses the new rollback_plan pathway. - Tests and behavior coverage were expanded via the new Behave feature and step implementations to validate state handling, events, and delegation. - Affected modules/components - src/cleveragents/infrastructure/events/types.py - src/cleveragents/application/services/plan_lifecycle_service.py - src/cleveragents/cli/commands/plan.py - PlanLifecycleService module docstring - features/plan_lifecycle_rollback.feature - features/steps/plan_lifecycle_rollback_steps.py ISSUES CLOSED: #3677	2026-06-03 03:22:59 -04:00
hurui200320	87a7ce35d7	feat(session): implement real LLM actor invocation in session tell ... Replace the M3 stub in `agents session tell` with real orchestrator actor invocation via `SessionWorkflow.tell()`. The stub echoed a canned `Acknowledged: ...` response without calling any LLM or actor; this commit wires up the full pipeline: Architecture: - New `SessionWorkflow` (Application layer) orchestrates LLM invocation for session tell. Accepts a `ProviderRegistry` and `ToolRegistry`; falls back to `FakeListLLM` when no provider is configured. - `LangChainSessionCaller` implements the `LLMCaller` protocol, building history-aware LangChain message lists from the session conversation and invoking the LLM via `ToolCallingRuntime.run_tool_loop()`. - `TellResult` (Pydantic BaseModel) carries the assistant response plus token usage (input_tokens, output_tokens, cost_usd, duration_ms). Domain / service layer: - `SessionActorNotConfiguredError` added to the session domain model; raised when tell is invoked with no actor on the session and no `--actor` override. Clear message; CLI exits with code 1. - `SessionService.get_messages()` abstract method added (+ implementation in `PersistentSessionService`) to load ordered message history. A2A facade: - `A2aLocalFacade` gains `message/send` and `message/stream` standard A2A operation handlers that route to `SessionWorkflow.tell()`. Total supported operations count: 42 → 44. CLI: - `session tell` command delegates to `_build_session_workflow()` (patchable factory) instead of directly calling `SessionService`. - Non-streaming: `SessionWorkflow.tell()` returns `TellResult`; output includes a Usage panel (Rich/Plain) or a `usage` object (JSON/YAML). - Streaming: `SessionWorkflow.tell_stream()` yields tokens; CLI prints them via `console.print` (not raw `sys.stdout.write`). - Token usage recorded via `SessionService.update_token_usage()` in both paths. Tests: - New Behave feature `session_tell_llm.feature` (4 scenarios): real LLM response persisted; streaming yields tokens; no-actor exits code 1; `--actor` override resolves correctly. - New Robot suite `session_tell_llm.robot` (4 tests): end-to-end with stub LLM injected via monkey-patching `_resolve_llm`. - Updated all existing `session tell` test steps to patch `_build_session_workflow` returning a mock `TellResult` so tests remain isolated from the LLM layer. - Updated operation-count assertions (42 → 44) in `a2a_cli_facade_integration`, `consolidated_misc`, `m6_autonomy_acceptance` feature files and steps. Cycle 7 (Review ID 8088) fixes: - Blocker 4: Split `session_workflow.py` (was 801 lines) into two files: `session_workflow.py` (467 lines) and `session_caller.py` (318 lines). Extracted: `LangChainSessionCaller`, `extract_content`, `extract_token_usage`, `estimate_cost`, `estimate_tokens`, `history_to_langchain_messages`, and stub classes. - Blocker 5: Route CLI streaming path through `_facade_dispatch("message/stream", ...)` instead of directly calling `workflow.tell_stream()`. The facade's `_handle_message_stream` falls back to non-streaming with `streamed: false` (acceptable for this milestone per the spec). - Updated streaming test assertion to validate full response presence (no longer checks token-by-token word positions, since the facade fallback returns a complete message). ISSUES CLOSED: #5784	2026-05-11 04:39:29 +00:00
freemo	9c6d69153e	fix(a2a): rename A2aRequest/A2aResponse fields to comply with JSON-RPC 2.0 wire format ... Rewrites the A2aRequest and A2aResponse Pydantic models to use the field names mandated by the JSON-RPC 2.0 specification, fixing a fundamental protocol compliance issue that prevented external A2A-compliant clients from communicating with the server. Changes: - A2aRequest: a2a_version→jsonrpc (fixed '2.0'), request_id→id, operation→method; auth field removed (not in JSON-RPC 2.0) - A2aResponse: a2a_version→jsonrpc, request_id→id, status+data→result (success path), timing_ms removed; added _result_xor_error validator enforcing mutual exclusion of result and error fields - A2aLocalFacade.dispatch(): updated to use request.method, request.id, result=data, error=A2aErrorDetail(...) - A2aHttpTransport.send(): updated to use request.method - CLI call sites (session.py, plan.py): updated A2aRequest(method=...) and response.result / response.error field access - All existing A2A Behave step files updated to new field names - New 35-scenario Behave feature (a2a_jsonrpc_wire_format.feature) covering serialisation, deserialisation, validation, and facade dispatch - New 7-test Robot Framework suite (a2a_jsonrpc_wire_format.robot) for end-to-end wire format verification ISSUES CLOSED: #1501	2026-04-03 00:29:46 +00:00
freemo	ad98d41d61	feat(a2a): implement _cleveragents/ extension method routing ... Add spec-aligned _cleveragents/ prefixed extension method routing to the A2A local facade per ADR-047. The facade now supports 42 total operations: 31 new extension methods across 6 families plus 11 legacy proprietary names retained for backward compatibility. Extension method families implemented: - _cleveragents/plan/* (13 methods): use, execute, apply, cancel, status, tree, explain, correct, diff, artifacts, prompt, rollback, list. Plan operations delegate to PlanLifecycleService when wired; new operations (cancel, tree, explain, correct, artifacts, prompt, rollback, list) have stub handlers returning safe defaults. - _cleveragents/registry/* (6 methods): tool/list, resource/list, actor/list, skill/list, action/list, project/list. Tool and resource list delegate to existing services; entity lists are stubs. - _cleveragents/context/* (4 methods): show (delegates to existing handler), inspect, simulate, set (stubs). - _cleveragents/health/* (2 methods): check, diagnostics/run. - _cleveragents/sync/* (3 methods): pull, push, status (stubs). - _cleveragents/namespace/* (3 methods): list, show, members (stubs). Legacy proprietary names (session.create, plan.create, etc.) continue to work via the same handler map, marked deprecated. Updated existing test assertions for the expanded operation count (11 -> 42). ISSUES CLOSED: #876	2026-03-18 17:33:30 +00:00
freemo	ec0b7631d0	refactor(a2a): rename ACP module and symbols to A2A standard ... Renamed src/cleveragents/acp/ to src/cleveragents/a2a/ and all 13 Acp* classes to A2a* per ADR-047 (A2A Standard Adoption). Updated all imports, structlog event names (acp.* → a2a.*), field names (acp_version → a2a_version), and test references across the entire codebase. This is a cosmetic rename only — no behavioral changes. ISSUES CLOSED: #688	2026-03-12 14:38:57 +00:00