claude-code-guide
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides templates and guidance for
CLAUDE.mdfiles, which define project-specific rules for AI agents. This documentation identifies a surface where external instructions can influence agent behavior. - Ingestion points: Instructions for agents to read and update project-level
CLAUDE.mdfiles. - Boundary markers: Recommends iterative verification of diffs and inspecting resulting instructions before deployment.
- Capability inventory: The skill itself does not invoke tools or perform operations; it provides documentation for external tool configuration.
- Sanitization: Mentions keeping secrets out of tracked instructions and verifying environment-specific facts.
- [EXTERNAL_DOWNLOADS]: The skill includes standard reference links to external resources for user assistance.
- Provides a link to official documentation for memory features.
- References a community-maintained guide hosted on GitHub.
Audit Metadata