The Agent Skills Directory

[COMMAND_EXECUTION]: The skill leverages the Bash tool to execute a wide array of security-focused command-line utilities. This includes reconnaissance tools (nmap, dig, subfinder), SAST scanners (semgrep, bandit, gosec), secret scanners (gitleaks, trufflehog), and infrastructure audit tools (checkov, tfsec, trivy). These operations are consistent with the skill's primary purpose as a security reviewer.
[EXTERNAL_DOWNLOADS]: The reference guides provide instructions for installing numerous security packages and utilities via standard package managers including npm, pip, go, and brew. It also references a well-known certificate transparency service (crt.sh) for reconnaissance purposes.
[PROMPT_INJECTION]: The skill is intended to process untrusted source code and infrastructure configurations for security auditing. This inherent functionality creates an indirect prompt injection surface, where malicious instructions embedded in the files being audited could attempt to influence the agent's behavior. The skill includes instructions for the agent to maintain focus on the audit scope and use specific testing guidelines to mitigate risks.
[DATA_EXFILTRATION]: While the skill includes tools and regex patterns specifically designed to locate hardcoded secrets, API keys, and private tokens within a codebase, there is no evidence of logic designed to exfiltrate this information to unauthorized remote servers. The patterns are provided as legitimate audit checks.

security-reviewer