What It Does
Check whether an agent run plan has budget caps, timeout controls, retry limits, and audit evidence before execution. Assess a redacted agent-run control plan for spend caps, timeout controls, retry limits, and evidence logging before a long-running or paid API workflow is approved. The skill runs offline against one redacted JSON evidence file and returns JSON with a PASS, REVIEW, or BLOCKED decision, blocking and review codes, and `mutated_files`.
Use Cases
- Teams preparing review-ready evidence before operational handoff.
- Maintainers who need deterministic local checks without contacting live systems.
- Operators documenting risks, gaps, owners, and next actions for release governance.
Requirements
- Python 3 standard library only.
- No API keys required.
- OpenClaw-compatible local skill workspace with redacted evidence notes.
Price
$4.99 one-time purchase.