AI Chaos Index: 58/100
HIGH
After Cap & Grow Phase 1: the bleeding has stopped. Domain extraction started, frontend cut by 84%. But the structural debt is far from resolved.
Delta: -29 points. Structural fragility reduced. Structural debt remains.
What Changed in Phase 1
Cap & Grow is our incremental remediation methodology: instead of rewriting everything, we “cap” the existing monolith (isolate it, stop adding to it) and “grow” new, properly structured domains alongside it.
| Metric | Before | After Phase 1 |
|---|---|---|
| Total files | 8 | 32 |
| Largest frontend file | 1,084 LOC | 174 LOC |
| Domains extracted | 0 | 1 (billing) |
| Test files | 0 | 2 |
| CI/CD | None | None |
| Architecture docs | None | slice.spec.md present |
| Legacy code | — | Isolated in legacy/ folder |
Root Cause Severity Breakdown
page.tsx reduced from 1,084 to 174 LOC. Billing domain extracted with proper layers.
Domain boundary established. Legacy bridge adapter provides intentional coupling.
README added. Slice specs document architecture. But missing .env.example, legacy/ confusion.
2 test files for billing domain. No frontend tests. Only 1 of 3+ domains covered.
Still no CI/CD. Tests exist but not automated. No pre-commit hooks.
Top Findings
Legacy Monolith Still Present (legacy/ folder)
The original monolithic code (page.tsx 1,084 LOC, api.py 619 LOC) is still present in the legacy/ folder. It creates confusion about which code is active.
Risk: Developers may accidentally modify legacy code. The legacy folder doubles cognitive load.
Only 1 Domain Extracted (Billing)
Cap & Grow has extracted only the billing domain. User management, usage tracking, and dashboard still live in the monolithic page.tsx.
Risk: New features in non-extracted domains still carry full monolithic risk.
No CI/CD Pipeline
Despite having tests for billing, there is no automated pipeline. Tests must be executed manually — and are easily skipped.
Risk: Tests without automation provide false security. Regressions still reach production.
Partial Test Coverage (~15-20%)
Billing domain logic is tested, but no frontend tests and no tests for other domains.
Risk: Billing is protected, all other areas remain vulnerable to regressions.
Missing .env.example
No environment variable template. Developers must discover configuration by reading code.
Risk: New developer setup is error-prone. Production deploys may fail from missing env vars.
Why Is It Still “High”?
Phase 1 focuses exclusively on structural extraction — isolating domains and stopping the bleeding. It does not add CI/CD, comprehensive tests, or deployment safety nets. Those come in Phase 2–3.
Think of it as emergency surgery: we stabilized the patient, but recovery takes more work. The score of 58 is honest — and that honesty is deliberate.
Cap & Grow Phase 1: Structure only (domain extraction, legacy isolation, architecture docs)
Phase 2: Test infrastructure, CI/CD pipeline, deployment safety
Phase 3: Performance optimization, monitoring, production hardening
Download the Sample Report
See the full Quick Scan report for this mid-remediation state.
Download PDF Report (ACI 58)PDF • 5 pages • AI Chaos Index Quick Scan Report
Stuck in the middle of a rescue?
A Quick Scan tells you exactly where you are — and what comes next.