grimoire-review
Pass
Audited by Gen Agent Trust Hub on Nov 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data from external code repositories, such as pull requests and diffs, which could contain malicious instructions aimed at overriding the agent's behavior.
- Ingestion points: The skill specifically ingests diffs, pull requests, and untracked files for review as defined in
SKILL.md. - Boundary markers: The instructions do not specify any delimiters or explicit boundary markers to separate the untrusted data from the agent's internal instructions.
- Capability inventory: While primarily read-only,
SKILL.mdacknowledges that the agent may run commands (e.g., tests, build scripts) if authorized by the user, which could be manipulated by injected instructions. - Sanitization: There are no instructions for sanitizing, escaping, or filtering potential instructions embedded within code comments or repository metadata.
- [NO_CODE]: The skill consists entirely of markdown documentation and guidance files (
SKILL.mdand thereferences/directory) and does not include any executable scripts, binaries, or automated installation steps.
Audit Metadata