design-to-code

What problem does it solve? When AI coding agents work from design documents, they repeatedly re-read external docs, burning tokens and drifting from the intended design. This Skill converts design documents into code skeletons (files, classes, methods) with contract-level comments embedded in place, so the code itself becomes the implementation guide. ## Core Features & Use Cases - Skeleton Generation with Contract Comments: Creates files, classes, method signatures, and fields directly from design docs, embedding parameters, return values, exceptions, and expected behavior as comments with design-doc anchors. - Batched Parallel Generation: Sorts sub-requirements by dependency DAG, splits large projects into batches, and dispatches order-independent batches in parallel via task-dispatch. - Architecture & Pattern Evaluation: Before generating, evaluates whether design patterns (strategy, factory, layered architecture) fit the design and confirms the approach with the user. - Safe Modification of Existing Code: Marks change points with TODO annotations (add/modify/delete/refactor) while preserving existing implementations. - Use Case: After finishing a design document for a user system with 5 sub-requirements, run this Skill to produce a full Python skeleton where every method carries its contract, then hand off to code-implement for filling in logic. ## Quick Start Generate a code skeleton from the design documents in my .requirements directory, embedding contract comments into each method.