Capture structured evidence of agent run outcomes for hook threshold tuning and quality trend tracking.
| Field | Description |
|---|---|
| Task type | Feature, bugfix, refactor, docs, config |
| First-pass result | Pass or fail before reviewer handoff |
| Retry count | Number of fix-validate cycles before acceptance |
| Hooks fired | Which PreToolUse hooks triggered and whether correctly |
| False positives | Hooks that denied valid work (note which and why) |
| False negatives |
| Quality problems that reached review missed by hooks |
| Files changed | Count of touched files |
| Line delta | Net lines added/removed |