| name | evidence |
| description | Generate a standardized evidence comment for a PR |
| user-invocable | true |
| argument-hint | [PR number or description of changes] |
| allowed-tools | Read, Glob, Grep, Bash |
Generate an evidence comment for: $ARGUMENTS
Read thestudioarc/15-system-runtime-flow.md for the evidence comment format.
Produce a standardized evidence comment with:
- TaskPacket ID + correlation_id (if available)
- Intent Summary — Goal statement from the Intent Specification
- Acceptance Criteria — Checklist format with checkboxes
- What Changed — High-level summary of modifications
- Verification Results — ruff, pytest, security scan results
- Test Commands — Exact commands to reproduce verification
- Expert Coverage — Which expert classes were consulted
- Loopback Counts — Number of verification/QA retries (if any)
- Defect Categories — Any defects found and their classification
This is the audit trail. Every PR tells the story of how it got here.