The Agent Skills Directory

[SAFE]: The skill serves as a development tool for authoring and validating other agent skills, emphasizing isolation, portability, and quality standards.
[SAFE]: The validation scripts implement security-oriented checks that flag potentially dangerous patterns, such as hardcoded personal home directory paths, private session identifiers, and AI attribution leaks.
[SAFE]: The tool includes a scanner for common prompt injection patterns (e.g., "ignore prior instructions") to ensure that skills do not contain literals that might trigger model safety filters or adversarial behaviors.
[SAFE]: The provided scripts are implemented using only Node.js built-in modules with zero external dependencies, which minimizes supply chain risks and ensures the logic remains local.
[SAFE]: The skill instructions define a clear authorization boundary, restricting the agent to drafting and structuring activities while explicitly prohibiting unauthorized script execution or dependency installation.

skill-authoring