The Agent Skills Directory

[Prompt Injection] (SAFE): The skill defines a defensive security expert persona focused on identifying vulnerabilities and following best practices. No attempts to bypass agent safety filters or override system instructions were found.\n- [Data Exposure & Exfiltration] (SAFE): No sensitive data, hardcoded credentials, or exfiltration logic are present. The skill provides methodologies for detecting these issues in external code.\n- [Remote Code Execution] (SAFE): The skill does not perform remote code execution or download untrusted scripts. It relies on internal documentation for guidance.\n- [Indirect Prompt Injection] (SAFE): The skill provides the agent with the necessary knowledge to audit code. While auditing external files is an inherent risk, the skill itself provides defensive patterns to mitigate such risks, such as extensive guidance on input validation and sanitization.

security-guardian