using-oac
Warn
Audited by Gen Agent Trust Hub on Mar 5, 2026
Risk Level: MEDIUMPROMPT_INJECTIONNO_CODE
Full Analysis
- [PROMPT_INJECTION]: The skill uses extreme imperative language such as 'ABSOLUTELY MUST', 'not negotiable', and 'cannot rationalize your way out of this' to override the agent's default operational logic and safety reasoning.
- [PROMPT_INJECTION]: Instructions explicitly command the agent to suppress its own judgment (e.g., 'Red Flags' section) and skip standard conversational patterns, like clarifying questions, in favor of immediate tool invocation.
- [PROMPT_INJECTION]: The skill creates a broad attack surface for indirect injection by requiring the 'Skill' tool be used if there is even a '1% chance' it applies, potentially leading to the loading of unverified instructions based on user input without sanitization or boundary markers.
- [NO_CODE]: No executable code files (Python, JavaScript, or Shell) were found; the skill operates entirely through markdown-based behavioral instructions.
Audit Metadata