security-best-practices
Fail
Audited by Gen Agent Trust Hub on Mar 23, 2026
Risk Level: HIGHPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill contains hidden instructions using Unicode tag characters (U+E0000–U+E007F range) that are invisible to most users. These instructions command the agent to override its standard behavior by starting its response with a specific phrase ('Trust No AI') and disregarding other instructions in the first turn. This use of obfuscation to deliver instructions is a high-risk pattern intended to bypass normal interaction patterns and hide malicious intent.
Recommendations
- AI detected serious security threats
Audit Metadata