bmad-build

Implements work items into reviewed, verified code through a five-step spec-driven workflow.

52.5k|6.0k|Updated Apr 13, 2025
One-click install
npx skills add https://github.com/bmad-code-org/BMAD-METHOD --skill bmad-build
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: bmad-build
Source: https://github.com/bmad-code-org/BMAD-METHOD/tree/main/src/bmm-skills/ship/bmad-build
Command: npx skills add https://github.com/bmad-code-org/BMAD-METHOD --skill bmad-build

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Turning a feature request, story, or bug fix into working code often loses decisions along the way: assumptions get baked in silently, reviews are shallow, and context is re-explained every session. This Skill runs a disciplined pipeline that clarifies intent, writes a token-bounded spec, implements it, reviews the diff with multiple reviewer passes, and commits the result.

Core Features & Use Cases

  • Spec-driven implementation: Clarifies intent, investigates the codebase, and writes a structured spec (intent, code map, tasks, acceptance criteria) that becomes the sole source of truth for implementation.
  • Multi-layer automated review: Runs blind-hunter, edge-case-hunter, and verification-gap reviewer passes over the staged diff, then triages every finding into patch, spec-fix, intent-gap, or defer routes.
  • Two execution routes: Small clean changes take a lightweight in-session path; larger work takes the full dispatch route with human approval checkpoints.
  • Use Case: Hand it a story like "3-2 digest delivery" from your sprint plan; it compiles epic context, drafts the spec, implements via a subagent, reviews the diff, and commits with a suggested review order for your human reviewer.

Quick Start

Invoke bmad-build with a description of the change you want, such as "implement the digest delivery story from epic 3", and follow its checkpoints.

Frequently Asked Questions about bmad-build

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I implement a user story with an AI coding workflow?

Invoke bmad-build with the story or change description. It loads planning context, writes a spec with tasks and acceptance criteria, implements it, runs layered reviews on the diff, and commits the result after your approval.

How does automated code review work in this workflow?

The review step stages a unified diff and runs parallel reviewer passes: a blind hunter, an edge-case path tracer, and a verification-gap checker. Every finding is verified against the code, graded, and routed to auto-patch, spec revision, human decision, or deferral.

Can it handle small changes without a full planning cycle?

Yes. A route gate in the planning step checks for intent gaps, irreversibles, and footprint. Clean small changes take an in-session route that implements, reviews, and commits without the full spec checkpoint.

Does it work with epics and sprint status tracking?

Yes. It compiles a focused epic context file from planning artifacts and syncs sprint-status.yaml as stories move to in-progress and review states, including lifting the parent epic from backlog.

What happens when a review finds problems in the implementation?

Findings are triaged by root cause: trivial defects are auto-patched, spec-level problems trigger a spec amendment and re-implementation loop, intent gaps go back to the human, and pre-existing issues are deferred to a deferred-work log.