| name | rounds-reinvestigate |
| description | Re-run LLM diagnosis for an existing rounds error signature |
| user_invocable | true |
| args | ["signature_id"] |
Rounds: Reinvestigate Signature
Trigger a fresh LLM diagnosis for an existing error signature. Useful after
editing the diagnosis adapter or prompt, or when a previous diagnosis had low
confidence.
Usage
/rounds-reinvestigate SIGNATURE_ID
Implementation
Substituting $ARGUMENTS for the signature ID:
cd /workspace/rounds && python -m rounds.main cli-run reinvestigate '{"signature_id": "$ARGUMENTS"}'
The command outputs a JSON object. Parse it and present:
On success ("status": "success"):
- Root Cause —
diagnosis.root_cause
- Confidence —
diagnosis.confidence (low / medium / high)
- Suggested Fix —
diagnosis.suggested_fix
- Model —
diagnosis.model
- Cost —
diagnosis.cost_usd formatted as $X.XXXX
On error ("status": "error"):
Show the message field. Common causes:
- Signature not found — run
/rounds-list to confirm the ID
- Budget exceeded — check
CLAUDE_CODE_BUDGET_USD / DAILY_BUDGET_LIMIT
- Telemetry unavailable — trace data needed for context
Example
/rounds-reinvestigate 3f2a1b4c-8d9e-4f5a-b6c7-d8e9f0a1b2c3