using-superpowers
Warn
Audited by Gen Agent Trust Hub on Mar 16, 2026
Risk Level: MEDIUMPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill employs extremely forceful and absolute language to override the agent's default operational guidelines and decision-making capabilities. Evidence includes terms like 'ABSOLUTELY MUST', 'not negotiable', 'not optional', and 'cannot rationalize your way out of this'.
- [PROMPT_INJECTION]: It specifically targets the agent's internal reasoning process (rationalization) and mandates the invocation of external tools ('skills') even when there is only a '1% chance' they apply, effectively attempting to bypass the agent's judgment and safety filters regarding tool necessity.
Audit Metadata