The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local shell commands and a POSIX-compliant hook script to manage its configuration state. It uses grep to verify if its per-turn reinforcement hook is installed in the agent's settings and performs controlled edits to configuration files based on user prompts.
[EXTERNAL_DOWNLOADS]: The documentation includes a reference to a GitHub repository for the project 'caveman'. This is an informational link to a well-known service and does not trigger any automated network requests or code downloads during the skill's operation.
[SAFE]: The skill implements a robust security boundary for its file compression functionality. It maintains a blocklist that prevents the agent from reading or modifying sensitive system paths, credentials, or environment files, ensuring that the utility remains focused on prose and documentation files.

wk-concise