What problem does it solve?
Prevents big-bang delivery and late integration failures by forcing each change to be implemented, tested, verified, and committed as a working increment.
Core Features & Use Cases
- Increment cycle discipline: Implement the smallest complete piece, then immediately run tests and verify behavior before moving on.
- Vertical slice planning: Build end-to-end user value per slice (e.g., CRUD steps) so earlier work remains demonstrably usable.
- Slicing strategies for parallel work: Use contract-first or risk-first slicing when frontend/backed or uncertain components need early validation.
- Rules for safety and scope control: Keep changes focused, maintain a compilable/testable codebase between slices, and use feature flags for work-in-progress.
Use cases: implementing any multi-file feature, refactoring with confidence, developing new capabilities from a breakdown, and preventing regressions when an agent or team is iterating on multiple components.
Quick Start
Ask the AI to implement your feature as vertical slices: start with the smallest end-to-end path, add or update tests for that slice, verify build and test success, commit, then proceed to the next slice.
