ccpe-system/workbench/analysis/review-agent-regression-2026-06-02/reports/README.md

Reports

Final comparison report target:

review-agent-regression-report.md

The final report should decide:

1. Whether original-kernel minimal Lite is better than full CCPE System Lite rewrite.
2. Whether Gemini and ChatGPT/Codex require separate production prompt variants.
3. Whether CCPE Forge migration rules should change for mature review agents.
4. Which prompt variant should remain candidate, draft, or active.