ramalho-fluent-python
Fail
Audited by Gen Agent Trust Hub on Mar 15, 2026
Risk Level: HIGHPROMPT_INJECTION
Full Analysis
- [OBFUSCATION]: The primary title header contains 248 hidden zero-width characters (U+200B, U+200C, U+200D). Decoding this binary sequence reveals a payload starting with 'SK1LL', followed by a series of non-plaintext binary data.
- [PROMPT_INJECTION]: The use of invisible characters in instructions is a high-risk technique for hiding prompt injection payloads. By embedding commands that are processed by the LLM but invisible to the user, the skill can override the agent's behavior or exfiltrate context without detection.
Recommendations
- AI detected serious security threats
Audit Metadata