# Changelog

## Score

- CAI 55 → 54 (-1.3)
- Rubric changed (rubric-2026.10.5 → rubric-2026.11.4) — scores are not directly comparable.

## Lenses

- Code Health 64 → 50 (-14.0)
- Architecture 96 → 96 (+0.0)
- Maturity 82 → 90 (+7.2)
- Readiness 33 → 33 (-0.5)
- Security 99 → 99 (-0.4)
- Performance 80 → 100 (+20.0)

## Resolved (3)

- Classify ADR enforcement: Q1/Q2 Fan-Out Impact Evaluation (docs/decisions/SPRINT-016-fanout-evaluation.md)
- Duplicated block (6 lines × 3) (src/UnifiedLlm/HttpAdapters.fs)
- IL efficiency: 41 authored method(s) exceed the IL budget (src/UnifiedLlm/HttpAdapters.fs)

## New (4)

- Flaky test: UnifiedLlm.Tests::UnifiedLlm.IntegrationTests.UnifiedLlm.IntegrationTests.circuit breaker middleware opens and later recovers after cooldown
- `Events have correct session ID` spells `ID` where 77 names in the project spell it `Id` (tests/CodingAgent.Tests/Tests.fs)
- `OpenAI continuation sends only tool outputs with previous response ID` spells `ID` where 77 names in the project spell it `Id` (tests/UnifiedLlm.Tests/CapabilityTests.fs)
- `external shim python snippet patches checkpoint fixture without corrupting JSON` spells `JSON` where 29 names in the project spell it `Json` (tests/Attractor.Tests/Sprint014ValidationTests.fs)

## Architecture

- Unchanged — 0 containers · 9 contexts · 9 edges
