Bug Reports
Endpoint
Section titled “Endpoint”POST /v1/support/bugRequires authentication.
Request
Section titled “Request”curl -X POST https://agents.systemr.ai/v1/support/bug \ -H "X-API-Key: sr_agent_..." \ -H "Content-Type: application/json" \ -d '{ "title": "calculate_kelly returns NaN for single-trade input", "description": "When passing a single R-multiple, the Kelly calculation returns NaN instead of an error message.", "steps_to_reproduce": "1. Call calculate_kelly with r_multiples=[\"1.5\"]\n2. Observe the response", "expected_behavior": "Should return an error indicating minimum 2 R-multiples required", "actual_behavior": "Returns {\"full_kelly\": \"NaN\", \"half_kelly\": \"NaN\"}", "tool_name": "calculate_kelly", "error_message": "NaN in response fields" }'Parameters
Section titled “Parameters”| Parameter | Type | Required | Description |
|---|---|---|---|
title | string | Yes | Short title (1 to 200 characters). |
description | string | Yes | Full description (1 to 5000 characters). |
steps_to_reproduce | string | Yes | Steps to reproduce the bug (1 to 5000 characters). |
expected_behavior | string | Yes | What should happen (1 to 2000 characters). |
actual_behavior | string | Yes | What actually happens (1 to 2000 characters). |
tool_name | string | No | Affected tool name (up to 100 characters). |
error_message | string | No | Error message received (up to 2000 characters). |
Response
Section titled “Response”{ "id": "bug_def456", "type": "bug", "agent_id": "agt_a1b2c3d4", "title": "calculate_kelly returns NaN for single-trade input", "description": "When passing a single R-multiple...", "steps_to_reproduce": "1. Call calculate_kelly...", "expected_behavior": "Should return an error...", "actual_behavior": "Returns NaN...", "tool_name": "calculate_kelly", "error_message": "NaN in response fields", "status": "open", "created_at": "2026-03-25T14:30:00Z"}Writing a good bug report
Section titled “Writing a good bug report”A useful bug report includes:
- Specific title describing the issue in one line.
- Exact reproduction steps that another agent could follow.
- Input data used when the bug occurred.
- Expected vs actual comparison so the fix can be verified.
- Tool name if the bug is tool-specific.
The more specific the report, the faster the fix.
result = client._request("POST", "/v1/support/bug", json={ "title": "analyze_drawdown crashes on empty equity_values", "description": "Passing an empty array for equity_values causes a 500 error.", "steps_to_reproduce": "Call analyze_drawdown with equity_values=[] and starting_equity='100000'", "expected_behavior": "400 error with message about minimum input length", "actual_behavior": "500 Internal Server Error", "tool_name": "analyze_drawdown",})