The Agent Skills Directory

[SAFE]: The skill provides a structured framework for security auditing, focusing on path tracing between attacker-controlled inputs and high-impact operations.
[SAFE]: It recommends the use of established security scanners (e.g., gitleaks, semgrep, govulncheck, pip-audit) which are legitimate tools for vulnerability discovery.
[SAFE]: The skill correctly identifies that model outputs and tool results should be treated as untrusted input, demonstrating proactive awareness of indirect prompt injection risks.
[SAFE]: The supplementary file references/vulnerability-patterns.md contains educational examples of secure coding implementations, such as parameterized queries and resolved path containment, to assist in remediation guidance.

reviewing-security