Understand the system the AI is building

The architecture, extracted from your code

Spaccy walks your project’s import graph and builds the C4 model — context, containers and components — as a versioned diagram in the repository. It isn’t a drawing someone made once in Miro that nobody ever updated again: it’s derived from today’s code and regenerated when it changes.

command/architecture-extractin your repodocs/architecture/model/

The domain model, with nobody drawing it

Entities, attributes and relationships are extracted from the database schema, the backend DTOs and the frontend interfaces. You see the domain the system actually implements — and the divergences between layers show up on their own.

command/domain-extractin your repodocs/architecture/model/domain/

The decisions, with the reasoning preserved

Every structural decision becomes a dated record with context, discarded alternatives and consequences. When someone asks why the system is built this way — six months from now, in a code review or in an audit — the answer is written down, not stuck in the memory of whoever was in the meeting.

in your repodocs/architecture/adr/

From requirement to test, with no gaps

Every feature carries the proposal → specification → review chain, and the traceability matrix links scenario, use case and acceptance criterion to the code and the test that cover them. Whatever’s left uncovered gets pointed out by name, instead of discovered in production.

command/traceability

From the diagram down to the line of code

In the cockpit the model is genuinely navigable: you drill down from context to container to component, open the UML detail of any element, and jump straight to the source file in the editor. Understanding a new part of the system stops being an excavation.