Python: [Bug]: Final output is empty after successful tool execution in declarative agent run

[hapiiiiGit]

Description

Description

I encountered an issue where the declarative agent successfully generated and executed the expected Python solution via python_executor, but the final assistant response was empty, resulting in an empty final_output.

The the tool execution is succeeded. However, after the tool result was returned to the model, the final chat_output contained an assistant message with empty contents, and the recorded final_output was "". :contentReference[oaicite:0]{index=0}

Environment / Config

• Agent config:

instructions: |
  You are a Python coding assistant. Your task is to complete the following Python code from the HumanEval benchmark.
  Use tools when they help, and give concise, evidence-aware answers.
  Note:
    1. Output only final valid Python code. The final answer must be executable Python code.

model:
  connection:
    kind: key
  options:
    temperature: 0.0
    maxOutputTokens: 512
    allowMultipleToolCalls: false
    chatToolMode: auto
tools:
  - kind: function
    name: python_executor
    description: Execute Python code in a short-lived subprocess.
    bindings:
      - name: python_executor

What Happened

• The model first generated a valid implementation and called python_executor
• The tool execution succeeded:

{
  "ok": true,
  "stdout": "1.0\n",
  "stderr": "",
  "returncode": 0,
  "timed_out": false
}

• After that, the model was called again with the tool result. The final response was:

{
  "role": "assistant",
  "contents": []
}

• The run ended with:

"finish_reason": "stop",
"final_output": ""

Expected Behavior

After successful tool execution, the agent should return the final valid Python code.

Code Sample

Error Messages / Stack Traces

Package Versions

agent-framework 1.3.0, agent-framework-declarative 1.0.0b260507

Python Version

Python 3.14.4

Additional Context

No response

[VIVEKJAINDTU]

Hi @hapiiiiGit , can you please add the error messages .