The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute industry-standard security scanners, including trufflehog, semgrep, bandit, trivy, and tfsec. These commands are provided as examples for the user to perform security audits on their infrastructure and codebases.
[DATA_EXPOSURE]: The skill includes instructions to search for hardcoded secrets and credentials within files using Grep and Read. This behavior is central to its stated purpose as a security audit tool and does not include any mechanisms for exfiltrating discovered data.
[PROMPT_INJECTION]: No attempts to bypass safety filters, override system instructions, or extract system prompts were detected. The instructions are focused on defensive security practices and remediation guidance.
[SAFE]: The skill effectively functions as an educational and operational resource for security best practices. The provided examples correctly contrast vulnerable code with secure alternatives for various common attack vectors like SQL Injection, Command Injection, and Path Traversal.

loom-security-audit