What problem does it solve?
It prevents “blind implementation” by converting your spec artifacts into a pre-implementation blueprint that preserves code conventions, file-by-file intent, and execution order before any code is written.
Core Features & Use Cases
- Generates
blueprint.md from spec artifacts: Builds a structured pre-implementation plan from tasks.md, spec.md, and plan.md, optionally using data-model.md and contracts/ when present.
- Supports doc-only or scaffolding workflows: Produces documentation-only output by default, or can scaffold new files with TODO stubs on disk (without TODOs inside the blueprint itself).
- Detects what’s new vs modified vs deleted: Classifies every task file path and handles “already complete” tasks via a pre-completed summary table to maintain 1:1 traceability with task IDs.
- Produces a developer-followable artifact: Ensures the blueprint contains complete, copy-pastable code blocks (or diff blocks for modified files) so the next step (
/speckit.implement) can execute from blueprint.md.
Quick Start
Run /speckit.blueprint.generate in doc-only mode to create a blueprint.md from your current spec artifacts, then review it before running /speckit.implement.