The Agent Skills Directory

[SAFE]: The skill serves as an educational and procedural resource for agent engineering, emphasizing reliability, safety, and structured communication between agent instances.
[COMMAND_EXECUTION]: The references include detailed documentation for programmatic invocation of the claude CLI via the Bash tool. It provides specific patterns for bypassing nesting guards using env -u and utilizing automation flags such as --dangerously-skip-permissions for non-interactive sessions.
[SAFE]: The utility script scripts/validate-agent.sh is a non-malicious local bash script designed to validate the frontmatter and structure of agent markdown files without external network calls or sensitive data access.

write-agent