The Agent Skills Directory

[COMMAND_EXECUTION]: The documentation describes the implementation of lifecycle hooks (PreToolUse, PostToolUse) that can execute shell commands from the scripts/ directory for validation and linting purposes.- [COMMAND_EXECUTION]: The skill provides guidance on configuring Model Context Protocol (MCP) servers, which involves executing shell commands (e.g., npx) to launch server processes.- [REMOTE_CODE_EXECUTION]: References are provided for using the WebFetch and WebSearch tools within research-oriented agents to interact with external web content.- [COMMAND_EXECUTION]: The documentation includes instructions for using the bypassPermissions mode, which allows automation pipelines to skip interactive permission checks for tool execution, noted as a high-authority configuration for trusted environments.

edit-tool