adaptive-mission
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a robust workflow for managing engineering missions with a focus on evidence-based verification. It uses structured markdown files (STATE.md, TRUTH.md, PLAN.md, REVIEW.md) to maintain process integrity and ensure that actions are grounded in verifiable facts rather than outdated assumptions.
- [SAFE]: The 'Guard' agent role is specifically designed to perform safety boundary checks before irreversible operations or interactions with real hardware, which is a positive security practice.
- [SAFE]: The methodology emphasizes '做与评分离' (separation of execution and review), requiring independent verification of results, which helps prevent confirmation bias and ensures that all claims of completion are backed by hard evidence like logs and data.
- [SAFE]: All external resource handling (links and documentation) follows a strict distillation process where facts are extracted, versioned, and verified, reducing the risk of acting on outdated or untrusted information.
Audit Metadata