What problem does it solve?
It identifies architectural friction that makes a codebase hard to understand, refactor, and test by finding shallow, tightly-coupled modules whose seams hide bugs.
Core Features & Use Cases
- AI-driven codebase exploration: navigates naturally to locate places where understanding requires bouncing across many files or where integration seams create risk.
- Deepening opportunities discovery: produces a numbered set of candidates that focus on improving testability by deepening shallow modules.
- Refactoring RFC generation: guides the user through framing constraints and then designs multiple interface options before drafting a GitHub issue RFC using the provided template.
- Testing-first architectural focus: classifies dependencies (in-process, local-substitutable, ports & adapters, external mock) to support boundary-focused tests instead of brittle internal layering.
Quick Start
Use it to improve architecture when you want the AI to explore your repository and propose deepening refactors as issue-ready RFCs.