claude-md-manager
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted local project files (e.g., manifest files, lockfiles, and READMEs) to generate project documentation (CLAUDE.md). This creates a surface for indirect prompt injection where malicious content in a project could influence the agent's behavior.
- Ingestion points: The scripts/analyze_project.py script reads multiple files including package.json, Cargo.toml, and README.md from the project directory.
- Boundary markers: There are no explicit instructions to use delimiters or 'ignore' blocks when incorporating data from these files into the generated CLAUDE.md.
- Capability inventory: The agent has the ability to read project files and write the CLAUDE.md file back to the project.
- Sanitization: No specific sanitization or validation of the ingested file content is performed before it is used to inform the documentation.
Audit Metadata