mad-decompose

To break a feature design document into shippable milestones, you can use automated decomposition to split large specs into dependency-ordered vertical slices. This eliminates manual gap analysis and prevents unreviewable large PRs by mapping integration points across entities and endpoints.

Parallel gap analysis in milestone planning uses specialized agent analysts to simultaneously map existing code, backlog items, and design features. This process identifies unplanned work and missing dependencies before grouping features into validated, shippable vertical slices.

Generating scoped idea files for vertical slice milestones involves batched creation of idea.md documents aligned with governance standards. These files are automatically scoped for each decomposed milestone, ready to feed directly into downstream specification workflows.

Yes, you can automatically decompose design documents covering multiple endpoints and data entities. The process handles complex specs by calibrating milestone sizes, validating dependency DAGs, and documenting integration points for iterative development.

The best way to avoid unreviewable large PRs in feature development is to decompose design documents into dependency-ordered, vertical-slice milestones. This approach automatically splits large features into 6-8 shippable milestones with scoped idea files for iterative development.

You should not use automated design doc decomposition for small features that do not cover multiple entities, endpoints, or infrastructure components. It is designed for large specifications requiring vertical slice splitting, DAG validation, and parallel code and backlog analysis.