The Agent Skills Directory

[SAFE]: The skill implements a robust security-first approach by defining explicit operational boundaries that forbid the agent from performing destructive Git operations, such as force-pushes or hard resets, ensuring the preservation of repository history.\n- [SAFE]: It includes a dedicated security classification framework that identifies and flags critical changes in authentication, cryptography, and secrets management, routing these findings to appropriate security review channels.\n- [SAFE]: All referenced external tools and utilities, such as gitleaks, trufflehog, and commitlint, are well-established industry-standard tools for maintaining code quality and security, used here to reinforce best practices.\n- [SAFE]: The skill incorporates detailed technical guidance on identifying and managing risks associated with AI-generated code, specifically highlighting the increased vulnerability profile and requiring enhanced human oversight and automated scanning.

guardian