cox-tooling-excellence
Fail
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: HIGHOBFUSCATIONPROMPT_INJECTION
Full Analysis
- [OBFUSCATION]: The file
SKILL.mdcontains hidden data using zero-width characters (U+200B, U+200C, U+200D) embedded in the primary header '# Russ Cox Style Guide'. This steganographic technique hides a payload that is invisible to human users but remains part of the text processed by the AI model. - [PROMPT_INJECTION]: The hidden payload embedded via zero-width characters constitutes a prompt injection attack. This method is used to deliver instructions that can override the agent's system prompt or bypass safety filters without being detected in a standard code review.
Recommendations
- AI detected serious security threats
Audit Metadata