Aryan SinghandClaude Opus 4.8 6b1c461a80 fix(agents): survive long overnight runs ...

- compact at memory-safe token budget (~20k) not the 131k window; gpt-oss
  hits oMLX's prefill memory guard ~47k tokens on a 24GB machine
- repair invalid_tool_calls (malformed JSON args) before echoing history —
  LangChain serializes the raw malformed string which oMLX 500s on
- retry invokes (4x) with an oMLX watchdog restart between attempts
- ALWAYS finalize: salvage picks from findings even after invoke failures

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016DvdRZkyCy9NJpHYDdPqmS

6b1c461a80 · 2026-07-02 12:22:29 -05:00

History

..

__init__.py

feat(agents): pydantic output schemas (Pick/AgentOutput/RunManifest) + tests

2026-07-01 16:29:29 -05:00

compaction.py

fix(agents): survive long overnight runs

2026-07-02 12:22:29 -05:00

graph.py

fix(agents): survive long overnight runs

2026-07-02 12:22:29 -05:00

schemas.py

feat(agents): pydantic output schemas (Pick/AgentOutput/RunManifest) + tests

2026-07-01 16:29:29 -05:00