project-skill-protocol
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes content from project files (overlays) that could theoretically contain malicious instructions if the repository itself were compromised.
- Ingestion points: Reads files from the
docs/project-protocols/directory. - Boundary markers: Uses Markdown and YAML frontmatter to structure overlay content.
- Capability inventory: File read/write access and interaction with other agent capabilities.
- Sanitization: Employs an 'Additive-only' filter that explicitly refuses any rules attempting to bypass framework safety gates, git discipline, or user-confirmation requirements. It also incorporates a mandatory 'Proposal Gate' where the user must review and approve all generated or modified rules before they are committed to disk.
- [DATA_EXPOSURE]: The skill implements a secure path derivation mechanism. It ignores potentially malicious links in the index file and instead constructs file paths based on a validated project root and slug. This design effectively mitigates path traversal risks that could otherwise be used to access sensitive files outside the intended directory.
Audit Metadata