The Agent Skills Directory

[COMMAND_EXECUTION]: The skill requests and uses Bash access to execute legitimate security scanning tools such as semgrep, bandit, gitleaks, and trivy. This is necessary for its primary function of identifying security vulnerabilities in a target codebase.
[EXTERNAL_DOWNLOADS]: The documentation references installation commands for well-known security tools from official package registries (npm, PyPI) and GitHub repositories belonging to recognized security vendors (e.g., Aqua Security, Semgrep, Gitleaks). These are documented neutrally as part of the tool setup instructions.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data (source code, logs, and configuration files) which could contain malicious instructions. This is an inherent risk for security analysis tools.
Ingestion points: The agent reads project files and command outputs using Read, Grep, and Glob tools.
Boundary markers: The skill provides a structured reporting workflow and template to maintain focus on the audit task.
Capability inventory: Use of Bash for scanning and Read for file inspection.
Sanitization: The skill does not implement internal sanitization but provides extensive documentation in references/vulnerability-patterns.md teaching the agent and user how to identify and prevent injection vulnerabilities in other applications.

security-reviewer