The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use CLI commands to manage configuration files.
Evidence: The agent is guided to run agentsync sync, agentsync add, agentsync update, and agentsync resolve to synchronize and scaffold the .ai/src/ directory.
[EXTERNAL_DOWNLOADS]: References the execution of external tools from well-known providers.
Evidence: The skill documentation suggests using npx -y @anthropic/mcp-playwright to set up Model Context Protocol (MCP) servers, utilizing a package from a well-known organization.
[SAFE]: The skill includes instructions and examples that promote security best practices.
Evidence: Provides examples of configuring tool permissions to deny access to sensitive files like .env and specifies path sanitization for the agentsync add command to prevent traversal issues.
[SAFE]: Documentation of dynamic context injection syntax is provided for educational purposes.
Evidence: The description of the !shell command syntax in SKILL.md is for authoring custom slash commands in supported tools and does not trigger execution within this skill's own prompt context.

agentsync