What problem does it solve? Codebases accumulate shallow modules, tight coupling, and leaky boundaries that make them hard to test and navigate, and generic refactoring advice often ignores the project's own documented decisions. This Skill finds architecture-deepening opportunities that are grounded in the project's own domain language and ADRs. ## Core Features & Use Cases - Documented-model grounding: Reads OMNIGRID-CONTEXT.md and omnigrid/artifacts/adrs/ so every proposal uses the project's glossary terms and respects locked architectural decisions. - Shallowness scanning: Detects tight coupling, leaky boundaries, shallow modules, test resistance, and AI-navigation friction, each anchored to a concrete file:line location. - Surgical deepening proposals: Produces a prioritized list of minimal, independently shippable refactors with the signal removed, the payoff, and the ADR relationship. - Use Case: After a codebase grows organically, run this Skill to get a prioritized list of consolidation and extraction moves that feed directly into the grill-with-docs → to-prd → to-issues feature flow. ## Quick Start Ask the agent to analyze this repository for architecture-deepening opportunities using the documented domain model and ADRs, and produce a prioritized refactoring list.
