docs-keeper
Pass
Audited by Gen Agent Trust Hub on Feb 20, 2026
Risk Level: SAFE
Full Analysis
- [Indirect Prompt Injection] (SAFE): The skill defines a surface where the agent processes project code and commits to generate documentation plans. 1. Ingestion points: Project root directory, code files, and commit history. 2. Boundary markers: Clearly defined human-maintained and agent-maintained file zones. 3. Capability inventory: File system read/write and symbolic link creation. 4. Sanitization: Relies on standard agent behavior and provided structural rules.
- [Command Execution] (SAFE): Suggests the use of the 'ln' command for mapping instruction files, which is a benign and standard development practice.
Audit Metadata