What problem does it solve? When AI jumps straight from a feature request to code, names drift from existing conventions, scope creeps mid-implementation, and no decision record survives. This Skill inserts a staged workflow between the idea and the code so every feature has a design document, a checklist, and an acceptance report. ## Core Features & Use Cases - Stage Routing: Inspects existing artifacts in codestable/features/ and tells you which sub-skill to trigger next (brainstorm, design, fastforward, implement, or acceptance) instead of guessing from verbal descriptions. - Four-Stage Pipeline: Coordinates brainstorm (optional triage), design (design.md + checklist.yaml), stepwise implementation, and acceptance closure, with human checkpoints between stages. - Fastforward Mode: Compresses design into four sections for small, clear requests, while blocking fastforward for cross-subsystem or terminology-risky work. - Use Case: You say "add a CSV export capability". The Skill checks the feature directory, finds no artifacts, and routes you to cs-feat-design (or cs-brainstorm if the idea is still fuzzy), producing a dated feature folder with design and checklist files. ## Quick Start Tell the AI "I want to add a new feature: CSV export for monitor data" and let it inspect the features directory to route you to the right sub-skill.
