What problem does it solve? Turning a raw feature request or story into reviewed, verified code normally requires constant human supervision across planning, coding, and review. This Skill runs one full unattended development iteration: it clarifies intent, writes a development-ready spec, implements it via a subagent, and runs layered code reviews with automated triage and repair loops. ## Core Features & Use Cases - Intent-to-Spec Planning: Resolves intent from prompts, epic stories, or existing spec files, compiles epic context from planning artifacts, and produces a structured spec with code maps, acceptance criteria, and verification commands. - Autonomous Implementation: Hands the spec to a context-free implementation subagent, then runs the spec's verification commands and audits edge-case matrix test coverage. - Multi-Layer Review with Triage: Runs parallel review layers (blind hunter, edge-case hunter, verification-gap, intent-alignment), classifies findings into intent_gap/bad_spec/patch/defer/reject, and loops repairs up to five iterations before committing. - Use Case: Point it at a story in stories.yaml or describe a feature; it produces a committed, reviewed change with a spec file recording the full run result, triage log, and deferred findings. ## Quick Start Invoke bmad-build-auto with a feature description or a spec folder and story id to run one unattended plan-implement-review iteration.
