using-superpowers
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEPROMPT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill employs aggressive and non-negotiable directives (e.g., 'ABSOLUTELY MUST', 'not negotiable', 'not optional', 'you cannot rationalize your way out of this') to override the agent's standard interaction patterns, such as asking clarifying questions or exploring context before tool invocation.
- [INDIRECT_PROMPT_INJECTION]: The skill mandates the loading of external content (skills) whenever there is a '1% chance' of relevance. This significantly lowers the barrier for the agent to ingest potentially untrusted or malicious instructions found in the workspace or provided via user-influenced filenames. Furthermore, it instructs the agent to 'follow skill exactly' once loaded, reducing the agent's ability to filter out malicious tasks.
Audit Metadata