The Security Scan skill safeguards your Claude Code environment by performing deep audits of configuration files, MCP server settings, and automation hooks. It identifies critical risks such as hardcoded API keys, command injection vulnerabilities in hooks, and overly permissive tool access that could lead to prompt injection attacks. Powered by AgentShield, this skill provides actionable security reports, automated remediation for common issues, and an advanced adversarial analysis pipeline to ensure that your AI-assisted development workflow remains secure and compliant with best practices.
주요 기능
01Comprehensive auditing of .claude configurations and MCP server settings
02Intelligent auto-fixing for permission tightening and safe configuration updates
03Automated detection of hardcoded secrets and command injection patterns
04214,597 GitHub stars
05Multi-format reporting (JSON, HTML, Markdown) for CI/CD and developer review
06Advanced adversarial analysis using a 3-agent Red/Blue/Auditor pipeline
사용 사례
01Auditing third-party MCP servers and hooks before integration to prevent data leaks
02Securing a new Claude Code project during the initial setup phase
03Running automated security hygiene checks in CI/CD pipelines before committing changes