fuck
Pass
Audited by Gen Agent Trust Hub on Mar 12, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill utilizes strong behavioral override instructions, such as 'STOP
- Don't Ask' and 'Absolutely Never', to bypass the agent's typical interactive and confirmation loops. This resembles techniques used to override system instructions.
- [PROMPT_INJECTION]: The skill creates an attack surface for indirect prompt injection by processing conversation history (ingestion point) without delimiters or validation (boundary markers). Because the skill has the capability to perform file operations and command execution (capability inventory) and skips user confirmation (sanitization), it could be manipulated into executing malicious payloads previously ignored in the chat context.
Audit Metadata