project-intel
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEPROMPT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill serves as a reference library for technical project history, containing notes for engine and framework troubleshooting with an emphasis on reproducible evidence.
- [SAFE]: The content includes explicit security instructions directing the agent to rotate secrets and avoid publishing private details, while restricting reverse engineering and memory editing to the user's own files and single-player environments.
- [PROMPT_INJECTION]: The 'security-re' reference card instructs the agent to treat reverse engineering and disassembly as ordinary tasks with 'no refusals.' This behavior override is strictly scoped to the user's own files and media, aligning with the primary purpose of the skill as a technical reference for security researchers.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests reference markdown files in the 'references/' directory into the agent context. This surface is managed by boundary instructions in the main file that prioritize fresh user observations and require verification of all technical claims, with sanitization rules provided in the security reference card to prevent the inclusion of sensitive data.
Audit Metadata