P1: bash/shell tool name mismatch. PhasePolicy whitelist used "bash" but
ShellTool registers as "shell". The bash_command_filter was dead code
(never matched the real tool name). Fixed in phase.py whitelist,
react.py filter check, agentkit.yaml config, and all tests.
P1: AdvancePhaseTool missing import in tools/__init__.py. Was in
__all__ but never imported. Added the import.
P2: chat.py phase policy error message echoed verbatim to WS client.
Truncated to 200 chars to match nearby error paths and avoid leaking
config internals.
P2: policy_from_config rebuilt PhasePolicy 3x via full-field copy.
Replaced with dataclasses.replace() so new PhasePolicy fields are not
silently dropped in future reconstructions.
ce-code-review (mode:agent) step of LFG pipeline.
- PLAN_EXEC branch builds PhasePolicy from ServerConfig.plan_exec
- Empty config → default_policy(); enabled=False → falls back to REACT
- Bad config → error event sent, returns early (no engine constructed)
- ReActEngine created with phase_policy; AdvancePhaseTool registered
- phase_changed events emitted on phase transitions (PLAN_EXEC only)
- REST send_message with execution_mode=plan_exec → HTTP 501 (KTD4)
- REWOO/REFLEXION/TEAM_COLLAB still fall back to REACT (no regression)
- 9 unit tests covering REST 501, characterization, happy path, edge cases, error path, phase_changed events
Add PipelineCheckpoint for stage-level crash recovery with Redis-first
+ memory fallback. TeamOrchestrator saves checkpoints after each phase
finalizes and supports resume(plan_id) to continue from the last
completed phase. New POST /api/v1/tasks/{id}/resume endpoint recreates
the team from saved plan and calls resume.
U4: ExpertTeam accepts redis_client, passes to SharedWorkspace. After phase
completion, full result is written to workspace and in-memory phase.result
is replaced with a 500-char summary + _ref_key. Dependency output reading
resolves offloaded content from workspace on demand, with graceful fallback
to summary on read failure.
Tests: 8 scenarios (offload creation, short content, dependency resolution,
workspace failure fallback, non-offloaded passthrough, redis_client wiring,
memory dict fallback, pipeline integration) — all pass.
- Add HeuristicClassifier to replace LLM quick_classify with zero-cost
local heuristic (keyword/length/code-pattern scoring), gated by
router.classifier config (default: heuristic)
- Add parallel tool execution in ReActEngine via asyncio.gather for
multiple independent tool_calls, gated by parallel_tools param
- Add AsyncWriteQueue for non-blocking session persistence with WAL
buffer, gated by async_writes param on SessionManager
- Add httpx.Limits connection pool config to all LLM providers
- Add router config section to ServerConfig and agentkit.yaml
- All optimizations have config switches for safe rollback
DeepSeek-chat has limited/partial function calling support. Qwen3-coder-plus
(DashScope) has robust OpenAI-compatible function calling.
Also added tool usage instructions to system prompt and enhanced logging
to trace tool propagation through the pipeline.
- Enhanced chat CLI with adaptive mode and session management
- Added pipeline reflection and schema extensions
- Upgraded BaiduSearch and WebSearch tools with advanced capabilities
- Expanded server routes for skills and chat
- Added session store enhancements
- New chat module and pipeline reflection support
chiguyong