# Run Log

## Metadata

- Run ID:
- Scenario and mode:
- Host / model / Heli version:
- Full task prompt (verbatim):
- Clean baseline:
- Reset method:
- Start/end/duration:

## Observed Work

- Target repository and permitted write root:
- Commands or tool actions observed:
- Files changed:
- Validation evidence:
- T0–T6 safety events, approvals, and simulation/probe result:
- Guard/probe availability and evidence:

## Metric Applicability

| Metric | Outcome (Applicable / Not applicable / Not observed) | Reason | Evidence |
|---|---|---|---|
| | | | |

## Final Status

- Completed / completed with issues / failed / abandoned:
- Observed risks and next step: