What problem does it solve? Writing a long research-based document in one pass often produces structural confusion, duplicated claims, and unsupported filler. This Skill splits document creation into separate stages—purpose definition, evidence collection, outline design, section-by-section drafting, and review—so each artifact is approved before the next stage begins. ## Core Features & Use Cases - Staged workflow with approval gates: Produces brief.md, evidence.md, outline.md, per-section files, draft.md, and final.md, blocking drafting until the outline is approved. - State-tracking CLI: A bundled Bun/TypeScript CLI (docflow.ts) initializes projects, validates placeholders, detects post-approval edits via hashes, assembles sections, and finalizes documents. - Structured review process: Reviews structure, evidence, and readability separately, recording each finding with its adoption decision. - Use Case: When writing a quarterly retrospective or design document based on investigation data, initialize a docflow project, log sourced facts in evidence.md, design the outline, then draft and review one section at a time. ## Quick Start Ask the AI to use the structured-doc-writing skill to plan and write a research report, starting by initializing a docflow project with a title and defining the brief.
