using-superpowers

Pass

Audited by Gen Agent Trust Hub on Apr 8, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill employs 'non-negotiable' and 'extremely important' markers to force a specific behavior (invoking the Skill tool) regardless of the agent's internal logic or the simplicity of the user request.
  • [PROMPT_INJECTION]: Instructions specifically command the agent to ignore its own reasoning process ("You cannot rationalize your way out of this") and to disregard typical conversational flow in favor of a rigid tool-first constraint.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 8, 2026, 09:27 PM
Security Audit — agent-trust-hub — using-superpowers