security-check

To audit Claude Code configurations for security risks, run a security-check against your project to cross-reference local and MCP setups against a bundled threat database, producing a structured vulnerability report in minutes.

A threat database detects insecure MCP server setups by analyzing MCP pins and checking for risky hooks, exposed credentials, and vulnerable memory prompts across local and project scopes to surface configuration security risks.

Yes, you can check MCP pins and agent inventories without external dependencies. The security-check reads a bundled threat-db.yaml file to analyze configurations and surface vulnerabilities across local and project scopes.

The best way to surface risky hooks and operational patterns is a phase-driven security audit that cross-checks project, local, and MCP configurations against threat-database entries to generate an actionable remediation report.

Yes, security-check works with both project-level and local-level Claude Code configurations. It applies a phase-driven audit to MCP server setups, hooks, and agent configurations across multiple scopes to identify vulnerabilities.

Your security audit report shows exposed secrets because the threat database actively checks memory prompts and configurations for credentials exposure, identifying insecure operational patterns that leak sensitive data across project scopes.