What problem does it solve? Solo builders lack the specialist review a real team provides, so designs reach build with hidden data gaps, uncovered process steps, and vaguely defined work. This Skill runs an iterative review loop between the design artifact and the build decision, refining screens, defining slices, and maintaining a backlog until enough slices reach Ready state to start building. ## Core Features & Use Cases - Multi-lens review: Applies PM, tech lead, and data lenses to every screen, with an optional four-agent parallel mode (UX, data, process, scope specialists) in Claude Code. - Slice definition and gating: Defines units of work with plain-language and technical descriptions, anchors, done criteria, quality contracts, and a strict Ready checklist. - Persistent backlog: Maintains docs/backlog.md with slice statuses, a data questions log, review rounds, and a decisions changelog readable by anyone outside the conversation. - Use Case: After a design sprint produces HTML screens, run a review round to classify findings, trigger research spikes, check process coverage against the to-be map, and promote the first coherent set of slices to Ready so build can begin. ## Quick Start Run a design review round on the screens in docs/design and update the backlog with any slices that are ready to build.
