跳到正文

Break Work into Stories and Track It

这部分内容暂未提供中文版本。

Use this page to turn a plan into stories you can build in one session and keep track of them. The path depends on the plan: a spec-backed epic gets Story Breakdown; a project with a PRD gets epics and stories, then bmad-sprint-planning.

PlanDo thisTracking artifact
One epic backed by SPEC.mdAsk bmad-spec for Story BreakdownOrdered stories.yaml beside SPEC.md
A project with a PRD (and UX or architecture)Run bmad-create-epics-and-stories, then bmad-sprint-planningEpic files plus sprint-status.yaml

For a spec-backed epic, stories.yaml is the whole tracking file. Build creates each story’s implementation record under the spec folder, and Finish an Epic reads stories.yaml as the inventory. No sprint-status file is involved.

For a project, bmad-create-epics-and-stories works with you as a product partner to turn the PRD’s requirements and the architecture’s decisions into epic files organized by user value, each story carrying acceptance criteria a developer can implement against. Everything from here on is about that path.

Run bmad-sprint-planning at the boundary between planning and implementation. Before any tracking exists, it judges the plan like a skeptical senior developer reading a handoff. It inventories whatever planning documents the project actually has — briefs, PRFAQs, PRDs, specs, UX output, architecture, epics — by reading them, not by filename. Then it asks one question: could a developer implement these epics without inventing decisions nothing records?

The verdict is PASS, CONCERNS, or FAIL. Concerns are listed and you choose whether to proceed. A fail stops with findings ordered by severity, each naming the skill that fixes it. A missing document type is only a finding if stories depend on it; a project with no UX document and no UI stories is fine.

Say “check implementation readiness” to run only the gate. The IR trigger on the Product Manager’s and Architect’s menus does the same.

After the gate passes, the same skill generates sprint-status.yaml. Build syncs story statuses into it, code review moves stories through review, and the retrospective appends action items to it.

Re-running generation is safe: finished work stays finished, action items and hand-written comments pass through, and a dry run reports drift without writing.

Say “show sprint status” to skip the gate and see where you are: counts by status, risk flags (a stale file, orphaned stories, stories waiting in review), open action items from retrospectives, and one recommended next action with its story key. There are no time estimates: status, risks, and next steps only.

The next action follows a fixed priority: resume in-progress work, review what is waiting, start the next ready story, start the first backlog story, run an open retrospective, or report done.

Say “validate sprint status” to check the file’s format without changing it. Say “fix sprint status” when the file is broken or has drifted from reality. The skill infers the true state first — from epic files, story files, and git history — and shows you one proposed state table. Nothing is written until you confirm it. Then it regenerates a clean file and validates it. Repair is the only path that can mark a story as less complete than it was, because it reflects confirmed reality.

Old names still work: bmad-check-implementation-readiness and bmad-sprint-status forward here. Move any _bmad/custom/bmad-sprint-status.toml overrides to bmad-sprint-planning.toml.

Implement each story with bmad-build, or with bmad-build-auto once the decisions are stable. When the epic’s stories are done, close it with Finish an Epic.