The Agent Skills Directory

[SAFE]: The skill is composed of reference documentation and design patterns for AI agents. It does not contain any executable code, shell commands, or network requests.
[SAFE]: It emphasizes security and safety best practices, explicitly instructing developers on how to implement guardrails, input validation, and execution sandboxes to prevent common vulnerabilities like prompt injection.
[SAFE]: The skill instructs the agent to read local project configuration files (.maestro.md, .maestro/decisions.jsonl) to gather necessary context, which is a standard operational procedure for this type of development utility.

agent-workflow