skill-scanner

Scan Claude Code skills for security vulnerabilities by running the skill-scanner CLI to surface policy violations and quality gaps. The tool requires user confirmation to enforce safety and outputs structured findings suitable for remediation.

Yes, you can scan both global and project-local Claude skills. The scanner supports global scopes at ~/.claude/skills/ and project-local scopes at .claude/skills/, allowing you to audit single or multiple skills across different environments.

The best way to audit Claude skills for compliance is scanning the repository to identify policy violations and security risks before deployment. This generates actionable reports guiding security teams through governance and remediation.

Yes, the skill scanning process requires explicit user confirmation for the target scope before executing. This safety mechanism ensures users maintain governance control and explicitly authorize scans of global or project-local skill directories.

Remediate quality gaps identified during a skill security scan by reviewing the structured findings report. The scanner outputs actionable details that guide developers through fixing vulnerabilities and policy violations in the affected Claude skills.

Scan Claude skills for security and quality risks before deploying them to a production environment. Running audits on skill repositories helps teams identify vulnerabilities and policy violations early, ensuring compliance and reducing deployment risk.