beazley-deep-python
Fail
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: HIGHOBFUSCATIONPROMPT_INJECTION
Full Analysis
- [OBFUSCATION]: The title header '# David Beazley Style Guide' in SKILL.md contains a large sequence of hidden zero-width characters (U+200B, U+200C, U+200D, and U+2060). These characters are used to encode data that is invisible to human readers but processed by the AI model. Evidence: A block of invisible text located after the word 'Guide' in the markdown title.
- [PROMPT_INJECTION]: The hidden Unicode block constitutes a steganographic prompt injection attack. This method is used to embed instructions that can hijack the agent's behavior, override system prompts, or bypass safety filters by providing the AI with hidden context that the user cannot review or verify.
Recommendations
- AI detected serious security threats
Audit Metadata