Devin AI 10384152d3 fix: prevent tools from being discarded when response_model is set (#4697) ...

When both native function calling (tools) and output_pydantic (response_model)
are enabled, InternalInstructor was intercepting the LLM call and creating its
own completion without passing tools. This caused agent tools to be silently
discarded.

The fix checks whether tools are present in params before routing to
InternalInstructor. When tools exist, the normal litellm.completion path is
used so the LLM can see and call the agent's tools. InternalInstructor is
still used when no tools are present (backward compatible).

Applied to all three response handlers:
- _handle_non_streaming_response
- _ahandle_non_streaming_response
- _handle_streaming_response

Co-Authored-By: João <joao@crewai.com>

2026-03-04 12:36:59 +00:00

..

crewai

fix: prevent tools from being discarded when response_model is set (#4697)

2026-03-04 12:36:59 +00:00

crewai-files

fix(deps): bump pypdf, urllib3 override, and dev dependencies for security fixes

2026-03-04 01:13:38 -05:00

crewai-tools

fix(deps): bump pypdf, urllib3 override, and dev dependencies for security fixes

2026-03-04 01:13:38 -05:00

devtools

feat: bump versions to 1.10.1a1

2026-02-27 09:44:47 -05:00