The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructional shell scripts for managing the lifecycle of security skill files. It uses standard commands like mkdir, cat, and ripgrep (rg) to create directories, write markdown files, and audit content for markers like TODO or FIXME. These operations are scoped to a local skills/ directory.
[DATA_EXPOSURE]: Uses the ripgrep utility to search through skill files for consistency and name collisions. This is a local read operation intended for quality control and documentation management.
[INDIRECT_PROMPT_INJECTION]: As a framework for generating instructions that other AI agents will execute, the skill represents an attack surface where untrusted data could be used to generate malicious instructions. However, the skill provides a structured methodology (Red-Green-Refactor) and specific validation steps to mitigate accidental errors or gaps in the generated content. (Severity: LOW).

writing-security-skills