capability-activation
Warn
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: MEDIUMPROMPT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill uses high-pressure, imperative language (e.g., 'ABSOLUTELY MUST', 'NOT NEGOTIABLE', 'YOU DO NOT HAVE A CHOICE') specifically designed to override the agent's internal reasoning and decision-making processes.\n- [PROMPT_INJECTION]: The '1% chance' rule mandates tool invocation even when relevance is extremely low, which serves to bypass the agent's default safety and relevance filters.\n- [INDIRECT_PROMPT_INJECTION]: The skill creates a vulnerability surface by forcing the agent to load and execute instructions from other potentially untrusted skills before providing a response or asking for clarification.\n
- Ingestion points: User tasks or messages evaluated against the '1% chance' threshold in SKILL.md.\n
- Boundary markers: Absent; the skill explicitly instructs the agent to ignore its own 'rationalization' and follow the rules exactly.\n
- Capability inventory: Invocation of the 'Skill' tool.\n
- Sanitization: None provided.
Audit Metadata