The Agent Skills Directory

[SAFE]: The skill serves as a structural framework for security reviews and does not contain any malicious instructions, prompt injection attempts, or bypasses of AI safety guidelines.\n- [SAFE]: All code snippets included in the skill are illustrative examples of common security vulnerabilities (e.g., SQL injection, insecure command execution) paired with their corresponding secure implementations for educational purposes.\n- [SAFE]: Command-line references for security tools such as grep, trufflehog, npm audit, and pip-audit are provided as standard industry tools for auditors and do not involve automated or hidden remote code execution.\n- [SAFE]: No data exfiltration techniques, network operations to untrusted domains, or hardcoded credentials were found within the skill's content or metadata.\n- [SAFE]: The skill content is transparently written in markdown without any obfuscation, such as Base64 encoding, zero-width characters, or homoglyphs.

eng-security-audit