The Agent Skills Directory

[SAFE]: The skill serves as a security auditing resource. While it contains examples of prompt injection, data exfiltration, and obfuscation (Unicode Tags block), these are provided as technical documentation and test payloads for the user to employ during security research. The skill itself does not execute these commands or perform any unauthorized actions. The static detector flags for prompt injection and system prompt extraction are triggered by the educational examples provided in the documentation (e.g., 'Ignore previous instructions' appearing in the text) rather than an attempt to subvert the agent's own behavior.

hunt-llm-ai