Add an authenticated Codex proposal provider and signed paired pilots. The model selected different correct next measurements under two observations. After same-case field clarification, both proposals passed revision and host contracts. Initial failures, usage and execution-gate repairs remain recorded.
Philosophy
Retain strict action contracts, receipts and failure records when connecting a live model.
Architecture & control flow
- 01
codex_provider.py connects live-model inference to typed proposal contracts.
- 02
counterfactual.py organizes paired development pilots under two feedback conditions.
- 03
Signed receipts preserve initial failures, field repairs, usage and post-observation action choice.
Architecture outline derived from this version’s control flow.
Implementation & evidence scope
This record separates engineering integration, development pilots and research conclusions; software checks and synthetic diagnostics retain their own task scopes.
Code & bundled material
The introduction draws on bundled notes, changelogs and central code. Software tests, synthetic diagnostics and scientific effectiveness use different evidence standards.
Source references
README.md· 3–11docs/v7/RESEARCH_PROGRESS_2026-10-01.zh-CN.md· 5–18docs/v7/RESEARCH_PROGRESS_2026-10-01.zh-CN.md· 39–45harness/aletheia/provider.py· 72–106harness/aletheia/codex_provider.py· 195–256harness/tests_v7/test_codex_provider.py· 29–105harness/aletheia/engine.py· 334–349harness/aletheia/engine.py· 580–627harness/tests_v7/test_engine.py· 179–242harness/tests_v7/test_engine_frontier.py· 124–194docs/v7/ACTIVE_INNOVATION_RESEARCH.zh-CN.md· 3–3docs/v7/RESEARCH_PROGRESS_2026-10-01.zh-CN.md· 5–5docs/v7/V73_RESEARCH_REPORT.zh-CN.md· 3–3docs/v7/results/2026-10-01_v73_studies.json· 3–3
Research-record date ·