dodds-testing-practices
Fail
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: HIGHOBFUSCATIONPROMPT_INJECTION
Full Analysis
- [OBFUSCATION]: The main header in
SKILL.mdcontains a significant block of hidden Unicode characters (U+200B, U+200C, U+200D, U+2060) placed between the title text and the header closing marks. This is a form of steganography designed to hide content from human readers while ensuring it remains in the context window for the AI model. - [PROMPT_INJECTION]: The presence of steganographic content in a skill file is a primary indicator of prompt injection. These hidden strings are typically used to deliver 'jailbreak' instructions, bypass safety guidelines, or override the agent's system prompt without the user's knowledge.
Recommendations
- AI detected serious security threats
Audit Metadata