local-security-check

To scan AI Skill files for security vulnerabilities, validate SKILL.md files to detect prompt injection, hardcoded credentials, and executable script risks. This process enforces compliance with Agent Skills Specification best practices and generates remediation steps.

Security checks for SKILL.md files include validating YAML frontmatter format, scanning for malicious directives and hard-coded credentials, and checking executable script risks. Local pre-commit validation enforces alignment with organizational guardrails.

Secret detection during a pull request code review scans Skill definition files to identify hardcoded credentials and data leakage risks. It validates YAML frontmatter and flags malicious directives to prevent supply chain risks.

Yes, you can use this for periodic compliance audits of existing AI Skills repositories. It audits all SKILL.md files to identify security risks, enforce alignment with the Agent Skills Specification, and generate a compliance report with remediation steps.

The best way to prevent data leakage in AI Agent Skill definition files is multi-layer security scanning. This detects prompt injection, hardcoded credentials, and executable script risks while validating format compliance within Skill creation workflows.

SKILL.md validation fails during local pre-commit checks when files contain YAML frontmatter format errors, malicious directives, hard-coded credentials, or executable script risks that violate Agent Skills Specification security best practices.