Tool Call Execution Problem with Structured Output

[AbdulHaseeb434000]

During the execution of agent with and without output_type parameter, I found this issue multiple times.
Without output_type specified, the agent executes tools accurately. However, when agent is restricted to structured output, it jumps to the answer and bypasses the tools execution. After multiple efforts, I found out that the behaviour still remained the same even when tool_use was changed to required in model_settings and mentioned clearly in instructions to use the tool.

[seratch]

When you have a specific output type for the whole agent run, the tool selection behavior could be changed especially when the tool's output is not compatible with the agent's final output type. In this case, your instructions in the agent may need to be more specific on when and how the attached tools should be used. With that, the agent may be able to more consistently and confidently use the tools. Hope this helps.

[github-actions]

This issue is stale because it has been open for 7 days with no activity.

[github-actions]

This issue was closed because it has been inactive for 3 days since being marked as stale.