The Agent Skills Directory

[COMMAND_EXECUTION]: The skill installs a shell script (enforce-subagent-model.sh) and configures a Cursor hook (hooks.json) to execute it. This script is used for policy enforcement to ensure subagents use a specific model and reasoning effort. The script uses jq to parse input safely and does not perform any network operations or access sensitive system files.
[SAFE]: The skill implements safety measures by requiring user confirmation before overwriting files or modifying .gitignore rules. It uses git check-ignore to respect project settings and provides clear reporting of all actions taken during the setup process.
[SAFE]: The skill uses a fixed set of local assets to assemble configuration files, preventing the injection of arbitrary content from memory or external sources.

rule-setup