deepwork_rules

To enforce file-change rules during AI agent sessions, you define rules as Markdown files with YAML frontmatter stored in the .deepwork/rules/ directory, which are then evaluated automatically at the Stop hook.

You can automatically trigger code reviews by defining rules with specific file patterns and detection modes. When matching files change, the rule evaluates at the session Stop hook and executes the designated prompt or command action.

Yes, you can govern AI documentation sync by storing automated governance rules as individual Markdown files with YAML frontmatter, specifying detection modes and actions to manage file modifications during AI runs.

Available detection modes for architectural governance rules include trigger, safety, set, and pair. These modes support executing either prompt-based or command-based actions when specified file patterns change during an AI session.

No, deepwork_rules requires zero external dependencies to manage AI workflows. It operates standalone, storing rules as Markdown frontmatter in the .deepwork/rules/ directory and evaluating them at the Stop hook.