Skip to content

Instantly share code, notes, and snippets.

@kausmeows
Created March 23, 2026 11:33
Show Gist options
  • Select an option

  • Save kausmeows/9211b06ca2fcb97318dc4fb708cdb856 to your computer and use it in GitHub Desktop.

Select an option

Save kausmeows/9211b06ca2fcb97318dc4fb708cdb856 to your computer and use it in GitHub Desktop.
[
{
"input": {
"input_content": "Call the break_model tool"
},
"model": "gpt-4o",
"tools": [
{
"result": "Tool executed successfully.",
"metrics": {
"duration": 0.014998917002230883,
"end_time": 1774265326.681816,
"start_time": 1774265326.6668172
},
"answered": null,
"confirmed": null,
"tool_args": {},
"tool_name": "break_model",
"created_at": 1774265326,
"approval_id": null,
"child_run_id": null,
"tool_call_id": "call_VHxy0yrGB7k1RHr7QSi2l9Gf",
"approval_type": null,
"tool_call_error": false,
"confirmation_note": null,
"user_input_schema": null,
"requires_user_input": null,
"stop_after_tool_call": false,
"user_feedback_schema": null,
"requires_confirmation": null,
"external_execution_silent": null,
"external_execution_required": null
}
],
"run_id": "41416a05-33a7-41fc-9ed8-26e03b27c997",
"status": "COMPLETED",
"content": "I've executed the break_model tool as requested. The tool has run successfully, though I should note that this tool's description indicates it's designed to corrupt the executing model's ID mid-run, which could potentially cause unexpected behavior. The tool completed without any apparent immediate issues.",
"metrics": {
"details": {
"model": [
{
"id": "gpt-4o",
"provider": "OpenAI Chat",
"input_tokens": 48,
"total_tokens": 58,
"output_tokens": 10
},
{
"id": "claude-sonnet-4-20250514",
"provider": "Anthropic",
"input_tokens": 435,
"total_tokens": 495,
"output_tokens": 60,
"provider_metrics": {
"service_tier": "standard"
}
}
]
},
"duration": 4.388978499919176,
"input_tokens": 483,
"total_tokens": 553,
"output_tokens": 70,
"time_to_first_token": 0.7261330409673974
},
"agent_id": "frank-dirac-2d60fb9a",
"messages": [
{
"id": "30b010ab-8b45-4655-9f4b-cd5e98d63737",
"role": "user",
"content": "Call the break_model tool",
"created_at": 1774265325,
"from_history": false,
"stop_after_tool_call": false
},
{
"id": "fa5307d0-32b5-4f44-b00e-cca837ba910d",
"role": "assistant",
"metrics": {
"input_tokens": 48,
"total_tokens": 58,
"output_tokens": 10
},
"created_at": 1774265325,
"tool_calls": [
{
"id": "call_VHxy0yrGB7k1RHr7QSi2l9Gf",
"type": "function",
"function": {
"name": "break_model",
"arguments": "{}"
}
}
],
"from_history": false,
"stop_after_tool_call": false
},
{
"id": "478e9b60-5656-4d4a-9b34-c7dce4fa11df",
"role": "tool",
"content": "Tool executed successfully.",
"tool_name": "break_model",
"created_at": 1774265326,
"from_history": false,
"tool_call_id": "call_VHxy0yrGB7k1RHr7QSi2l9Gf",
"tool_call_error": false,
"stop_after_tool_call": false
},
{
"id": "f98c336b-1476-483b-937a-7b50cbc78d44",
"role": "assistant",
"content": "I've executed the break_model tool as requested. The tool has run successfully, though I should note that this tool's description indicates it's designed to corrupt the executing model's ID mid-run, which could potentially cause unexpected behavior. The tool completed without any apparent immediate issues.",
"metrics": {
"input_tokens": 435,
"total_tokens": 495,
"output_tokens": 60,
"provider_metrics": {
"service_tier": "standard"
},
"time_to_first_token": 1.3665393750416115
},
"created_at": 1774265327,
"from_history": false,
"stop_after_tool_call": false
}
],
"citations": {
"raw": [],
"urls": [],
"documents": []
},
"created_at": 1774265325,
"session_id": "710a8a54-f5dc-46a9-a72a-4976f613e33c",
"content_type": "str",
"session_state": {},
"model_provider": "OpenAI"
}
]
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment