Skip to content

Roadmap

Released

v0.1 — Health Scoring & Provenance Capture

  • [x] Per-file health scores (1-10) combining churn, complexity, coupling, knowledge
  • [x] Hotspot detection ranked by ROI (churn x complexity x centrality x role)
  • [x] Co-change coupling detection from git history
  • [x] Knowledge risk / bus factor analysis
  • [x] ROI-aware ranking (core > test, central > leaf)
  • [x] Silent AI provenance capture via PostToolUse hooks
  • [x] Zero external dependencies
  • [x] Monorepo-safe filtering (lock files, configs, generated code excluded)
  • [x] Graceful degradation without git

v0.2 — Trend Tracking

  • [x] Health snapshots saved after each scan (once per calendar day)
  • [x] Trend comparison showing degrading/improving files with deltas
  • [x] Overall health trajectory
  • [x] Schema migration (v1 databases auto-upgrade)
  • [x] Marketplace distribution support

Planned

v0.3 — Drift Detection

  • [ ] Intent fingerprinting using code embeddings
  • [ ] Semantic drift measurement (original intent vs. current implementation)
  • [ ] /vitals:drift command to compare original purpose vs. current state
  • [ ] Drift score per file (0-100)

v0.4 — Decay Prediction

  • [ ] Per-module decay curves (exponential fit on health history)
  • [ ] Estimated liability dates ("this file becomes unmaintainable in ~47 days")
  • [ ] ROI-ranked refactoring plans with specific, actionable steps
  • [ ] Confidence scores on predictions

v1.0 — SaaS Dashboard

  • [ ] Team-level health views
  • [ ] Cross-repo decay comparison
  • [ ] Organizational technical debt portfolio
  • [ ] Benchmark data ("how does your team compare?")
  • [ ] CI/CD quality gates via MCP server

Contributing

Vitals is MIT licensed. Before contributing:

  • Keep the zero-dependency constraint (Python stdlib only)
  • No regex for file classification (structural signals only)
  • Test against a real repo with 100+ files
  • The report command must remain read-only (no side effects)