You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Apr 2, 2026. It is now read-only.
Integrate three new Supermodel analysis endpoints into the arch-docs pipeline so the generated site includes impact analysis, test coverage, and circular dependency data alongside the existing architecture documentation.
New Endpoints
POST /v1/analysis/impact
Blast radius for files and functions — direct dependents, transitive dependents, affected entry points, risk score
Accepts optional targets query param and diff file
Add impact data to entity pages: risk level pill, blast radius stats, affected functions list
POST /v1/analysis/test-coverage-map
Static test coverage via call graph reachability from test files (no test execution required)
Per-function tested/untested status with confidence levels, per-domain coverage percentages
Add coverage data to entity pages: tested/untested pill, test file links, coverage percentage on domain pages
POST /v1/analysis/circular-dependencies
Tarjan's SCC cycle detection at file level with severity scoring
Breaking suggestions for resolving cycles
Add cycle data to entity pages: "In Cycle" pill on files that participate, cycle detail with Mermaid diagrams
Implementation
Call all three endpoints in parallel with the existing /v1/graphs/supermodel call (same zip upload, separate idempotency keys)
Save response JSON alongside graph.json
Extend graph2md input or post-process markdown to inject analysis data into entity frontmatter and body sections
Add new taxonomies: test_coverage (Tested/Untested), impact_level (Low/Medium/High/Critical), dependency_health (Clean/In Cycle)
Add D3 visualizations for impact radius and cycle diagrams