renaissance-statistical-arbitrage
Fail
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: HIGHOBFUSCATIONPROMPT_INJECTION
Full Analysis
- [OBFUSCATION]: The main header of the skill contains a large payload of zero-width Unicode characters used for steganography. These characters are invisible to a human reader but are processed by the AI agent.
- Evidence: Approximately 230 hidden characters follow the word 'Guide' in the header
# Renaissance Technologies Style Guide. The payload uses a combination ofU+200B(Zero Width Space),U+200C(Zero Width Non-Joiner),U+200D(Zero Width Joiner), andU+2060(Word Joiner) to encode hidden content. - [PROMPT_INJECTION]: The presence of steganographic content in an instructional skill is a high-risk indicator of a prompt injection attempt. These hidden instructions are designed to influence the agent's behavior at runtime, potentially bypassing system prompts, safety guardrails, or instruction sets provided by the platform or the user.
Recommendations
- AI detected serious security threats
Audit Metadata