What problem does it solve? Architecture audits produce judgment-heavy findings that are hard to communicate to stakeholders who were not in the room. This Skill turns an existing components.json architectural model into a single self-contained HTML treemap that opens as a file:// page with no server, making component size, classification, metrics, and coupling edges legible at a glance. ## Core Features & Use Cases - Deterministic rendering: Converts a validated components.json manifest into a D3-based treemap with components sized by LOC and colored by core/seam/removable classification, refusing to render on schema failure. - Metric badges and provenance: Displays deterministic metrics (LOC, fan-in/out, cyclomatic complexity, 90-day churn, test ratio) alongside epistemic-source chips so readers can distinguish measured facts from judgment calls. - Coupling edge visualization: Renders statically detected edges as solid lines and audit-asserted edges as dashed lines, with import cycles detected via Tarjan SCC. - Use Case: After running a retrospective architecture audit on a Python codebase, render the synthesized components.json into a treemap.html and email it to architects and stakeholders to defend an OSS-core boundary or prioritize prune candidates. ## Quick Start Render the bundled golden fixture by running the render_treemap.py script against references/example/components.json and open the resulting HTML file in a browser.
