improving-architecturelisted

# Improving Architecture Analyze a codebase and surface architectural friction as **deepening opportunities** — refactors that turn shallow modules (a small amount of behaviour behind a large, leaky interface) into deep ones (a large amount of behaviour behind a small interface). The aim is testability and AI-navigability. The output is a **refactoring blueprint** emitted inline as Markdown. ## Read-only boundary This skill is strictly read-only analysis. It does NOT: - edit or refactor product code, or perform the refactoring it describes; - write `CONTEXT.md`, `CONTEXT-MAP.md`, or ADRs; - spawn agents or run multi-agent design exploration; - emit HTML or write any file — the blueprint is returned inline in the response. If asked to "just do the refactor" or "apply this," decline: this skill produces a **blueprint**, never executes one. Hand the blueprint to whatever workflow implements changes. ## Term discipline The artifact this skill emits is a **blueprint** — an analysis output, not an implementation plan. Keep the vocabulary distinct: a blueprint ranks and describes candidate refactors; turning one into an executable plan and applying it is a separate, downstream activity this skill does not perform. ## Glossary The skill uses two distinct vocabularies: - **Architecture terms** (module, interface, implementation, depth/deep/shallow, seam, adapter, leverage, locality) come from `${CLAUDE_PLUGIN_ROOT}/skills/_shared/LANGUAGE.md`. Use them exactly — do not drift