langchain==1.3.15
📦 langchainView on GitHub →
✨ 5 features🐛 12 fixes🔧 8 symbols
Summary
This release introduces new features like exposing `trace_policy` on `AgentMiddleware` and adding `reasoning_effort` as a chat model parameter. It also includes several bug fixes, such as preserving history on summarization failures and handling malformed structured-output responses.
✨ New Features
- Expose `trace_policy` on `AgentMiddleware`.
- Add `state_schema` parameter to `wrap_tool_call`.
- Filter internal middleware model calls from `messages` projection.
- Add LangSmith provider to `init_chat_model`.
- Add `reasoning_effort` as a standard chat model parameter.
🐛 Bug Fixes
- Preserve history on `SummarizationMiddleware` summary failure.
- Handle import error in `LLMToolEmulator` by `model`.
- Re-export `PIIMatch` from `middleware` package.
- Restrict narrowed `ToolStrategy` in bound tools.
- Handle malformed structured-output responses.
- Prevent orphaned `tool_calls` in `ToolCallLimitMiddleware` end behavior.
- Add aliases for bedrock mantle chat models.
- Clear stale `structured_response` between checkpointed turns.
- Stop HITL approval gates from silently failing open.
- Propagate model middleware control flow.
- Update Anthropic config test.
- Regression test for shell tool + checkpointer msgpack error.