# Changelog

## Baseline

- First survey — no prior run to compare against. CAI 43.7.

## Lenses

- Code Health 58.5
- Architecture 85.5
- Maturity 64.2
- Readiness 19.0
- Security 62.5
- Accessibility 86.0
- Performance 100.0

## Changes since last survey

- 300 commits — 288 feature/other, 12 fixes

## By area

- experiments/runs — 91 commits
- research/workflow-dev — 42 commits
- site/docs — 23 commits
- research/questions-cross — 21 commits
- research/questions-claude — 18 commits
- site/src — 18 commits
- (root) — 15 commits
- experiments/docker — 12 commits
- .claude/skills — 10 commits
- experiments/workflows — 7 commits
- research/questions-pi — 6 commits
- (repo) — 4 commits
- experiments/compute-cost.py — 3 commits
- todos_and_ideas/1-IN — 3 commits
- todos_and_ideas/1-IN.md — 3 commits
- experiments/measure-suite-transitions.py — 2 commits
- experiments/stacks — 2 commits
- research/kata-design — 2 commits
- research/kata-dev — 2 commits
- todos_and_ideas/2-ACTIVE — 2 commits

## Notable commits

- fix: Add completed Fable 5 runs (claim-office, v6.1-hybrid-testlist-scope-fix)
- fix: Fix mutation testing for CLI-driven suites, and fill RQ-current-models-ptdd
- fix: Fix three classifier defects the TCR/subagent/hybrid smoke runs exposed
- fix: fix(cost): one list-price source for every pi run, and real tariffs for the codex route
- fix: fix(export): expose Cursor workflow as exact-coding skill
- fix: fix(export): ship the two optional skills in every SOL subtree
- fix: fix(harness): gpt-6-astra was never declared in models.json
- fix: fix(parser): refactorings_applied auf CC aus Textmarkern statt Phasenquelle
- fix: fix(pi): kumulative Streaming-Events beim Transkript-Extrakt verwerfen
- fix: fix(pipeline): env-fault guard, container reanalysis, correct test counting
- fix: fix(rq): correctness on claim-office follows the spec style, not a test phase
- fix: fix(tooling): spot-check jq read the wrong block, and tabulate was undeclared
- change: Add /publish-site, and pull the host prerequisites back into line
- change: Add Claim Office workshop material: scenarios in prose, design decisions, art prompts
- change: Add Java EXACT Coding workflow comparison
- change: Add Java mutation testing via PIT and score RQ-exact-coding-java
- change: Add PTDD v1.2 verification split and open RQ-ptdd-verification-split
- change: Add RQ-current-models-ptdd: current models under the maintained PTDD workflow
- change: Add RQ-exact-coding-rust
- change: Add RQ-old-vs-new-exact-line-opus55 and fill its two retired-line cells
- …and 280 more

## Architecture

- 0 containers · 94 bounded contexts · 97 dependency edges (baseline)
