bos-concurrency-rust
Fail
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: HIGHOBFUSCATIONPROMPT_INJECTION
Full Analysis
- [OBFUSCATION]: The main H1 header in
SKILL.mdcontains a hidden sequence of zero-width characters (U+200B and U+200C) separated by zero-width joiners (U+200D). This form of steganography is designed to hide content from human review while remaining detectable and actionable by large language models.\n- [PROMPT_INJECTION]: The hidden Unicode characters in the title constitute a steganographic prompt injection. By embedding instructions that are invisible to the user but visible to the AI, the skill attempts to influence agent behavior covertly, which is an adversarial technique for overriding system prompts or bypassing safety guardrails.
Recommendations
- AI detected serious security threats
Audit Metadata