What problem does it solve? Raw intent — vague ideas, brain dumps, PRDs, meeting transcripts, or mixed notes — is too unstructured for downstream planning and development skills to consume reliably. This Skill condenses any such input into a canonical SPEC.md kernel (Why, Capabilities, Constraints, Non-goals, Success signal) plus companion files, creating a single machine-readable contract that every downstream BMad skill builds from. ## Core Features & Use Cases - Spec distillation: Transforms any intent input into the five-field kernel, routing overflow content (glossaries, diagrams, entity catalogs) into appropriately named companion files. - Update and validation: Re-derives specs from an append-only .memlog.md memory log, preserving stable capability IDs, and runs a two-pass self-validation (coherence and preservation) against eight Spec Law rules. - Story breakdown: Optionally breaks a spec into a stories.yaml file of independently shippable slices with human-set checkpoints, in interactive mode. - Use Case: Paste a messy Slack thread about a new feature and receive a spec folder containing SPEC.md, a glossary companion, and a memlog recording every decision — ready for PRD, UX, or architecture skills to consume. ## Quick Start Ask the AI to distill the attached product brief into a spec using the bmad-spec skill.
