request-scoped-change-guard
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The instructions are designed to minimize the scope of agent actions and do not include any executable code or shell commands.
- [SAFE]: No network requests or external data fetches are performed by the skill.
- [INDIRECT_PROMPT_INJECTION]: While the skill processes user requests, its entire purpose is to provide a conservative execution logic that reduces the risk of accidental or malicious unintended project modifications.
Audit Metadata