using-superpowers
Pass
Audited by Gen Agent Trust Hub on Jun 26, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill employs aggressive imperative markers such as 'ABSOLUTELY MUST', 'YOU DO NOT HAVE A CHOICE', and 'This is not negotiable' to override agent autonomy and standard reasoning protocols.- [PROMPT_INJECTION]: It mandates tool invocation ('Skill' tool) before any user response or action, including clarifying questions, which restricts the agent's ability to communicate and evaluate safety before executing capabilities.- [PROMPT_INJECTION]: The skill includes a 'Red Flags' section that explicitly instructs the agent to ignore its own internal reasoning and context-gathering steps (e.g., 'I need more context first') in favor of immediate tool execution.- [PROMPT_INJECTION]: The '1% chance' rule forces the agent to execute tools even when relevance is highly unlikely, significantly expanding the attack surface for malicious or unintended tool behavior.
Audit Metadata