writing-skills
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOWPROMPT_INJECTION
Full Analysis
- [Prompt Injection] (LOW): The skill promotes 'Authority' principles using keywords like 'YOU MUST' and 'No exceptions'. These are recognized patterns for bypassing AI safety filters or overriding instructions, even when used for benign process enforcement.
- [Metadata Poisoning] (LOW): The document cites a 2025 research paper (Meincke et al.) that is likely fictitious, representing a deceptive use of information to manipulate the agent's or user's perception of authority.
- [Data Exposure & Exfiltration] (SAFE): No network activity or sensitive file access was detected.
Audit Metadata