Research intake status

This directory contains reviewed extracts derived from the primary publications under reference_docs/. The PDFs are evidence inputs; the Markdown extracts are not executable defaults until their transcription and applicability are tested.

The superseded root-level research prompt and two raw agent reports were removed during the 26.9.9.1 release cleanup. One report was incomplete and the other contained non-durable tool citations from an older baseline. Their reviewed findings are represented by the source-pinned records below; the discarded raw intake is no longer retained.

Current reviewed extracts:

• local_evidence_inventory.md: SHA-256 and review/ rejection/pending status for evidence consolidated under reference_docs/.
• fhwa_box_inlets.md: corrected FHWA-HRT-06-138 Tables 11 and 12, exact local PDF locators, applicability limits, and the Appendix D configuration boundary. Implementation still requires typed geometry and inlet rules.
• nchrp734_coefficients.md: initial slipline and multi-barrel findings from NCHRP 734. The report's duplicated final Table 3-2 row label is resolved by the immediately preceding narrative but retained as a source typo; the separate HY-8 developer correction blocks use of the original embedded coefficients.
• fixture_candidates.md: the complete Bodhaine example audit, selected corrected-box fixtures, and the Austroads worked-example limitation.

The authoritative task status is in docs/work/2026-09-06-phase-0-to-10-remediation.md. CS-001 closed the research gate on 2026-09-09. The resulting deferred implementation work is tracked separately; research completion is not combined-solver validation.