Turn the current conversation context into a PRD and write it as a prd.md into the Overseer root. Use when the user wants to create a PRD from the current context.
Survey a project and interview the user divergently to surface a slate of rough candidate PRD shapes, written to the Overseer root as a staging doc. The breadth-first head of the authoring pipeline — run it before /overseer-grill-with-docs when you have a goal but don't yet know what PRDs it should break into. Use when the user wants to explore what features a project needs, or mentions "what PRDs", "shape the work", or "grill prep".
Interview the user relentlessly about a plan or design until reaching shared understanding, resolving each branch of the decision tree, while maintaining the domain docs (CONTEXT.md + docs/adr/) of the code repos in the session. Use when the user wants to stress-test a plan or design before writing a PRD, or mentions "grill me".
Test-driven development with red-green-refactor loop. Use when user wants to build features or fix bugs using TDD, mentions "red-green-refactor", wants integration tests, or asks for test-first development.
Break a PRD into independently-grabbable Issue files written into the PRD's folder under the Overseer root, using tracer-bullet vertical slices. Triage is folded in — each Issue is born routed (ready-for-agent / ready-for-human). Use when the user wants to convert a PRD or plan into Issues.
Approve a human-review Overseer Issue — merge its recorded worktree branch into the PRD feature branch, set the Issue status to done, and clean up the worktree. Use when the user has finished resolving a human-review Issue and wants to merge and complete it. This is the human-invoked twin of the reviewer's automatic clean-AI merge.