What this cut covers
- Gold sentinel: 34 works and 70 claims, independently recomputable from the accepted wave-0 authority.
- A1 researcher roster: 15 researchers with 15 affiliation-role claims and 53 representative-work claim occurrences. The roster establishes people and claims, not an A1 bibliographic-work count.
- A1↔C1 augmentation links: 11 claim-link occurrences across 8 distinct C1 identities are declared. 11 occurrences across 8 targets are active; 0 occurrences across 0 targets remain held until C1 identity admission.
- A1↔E1 work pointers: 8 pointer occurrences across 6 E1 targets are declared independently of researcher claims. 8 occurrences across 6 targets are active; 0 occurrences across 0 targets remain held.
- C1 OB/OT journals: 179 source works and 269 source claims were release-audited. Canonical identity resolution admitted 179 works carrying 269 claims; 0 works carrying 0 claims remain held, and 0 works carrying 0 claims were excluded.
- C1↔accepted-queue identity edges: 30 exact-DOI augmentation edges bind C1 rows to identities already present in the accepted 886-identifier queue. 30 are active and 0 remain held under the independent C1 lookup gate. The queue is identity evidence, not the publication base.
- B1 management canon: 110 corpus-evidenced canon candidates carrying 220 invocation-evidence statements were release-audited. Identity resolution admitted 74 works carrying 148 invocation statements; 36 works carrying 72 statements remain held, and 0 works carrying 0 statements were excluded. Canon entries are management ancestry, not AI papers; their own source-argument claims remain separate from invocation evidence.
- E1 SSRN backfill: 180 source works and 257 source claims were release-audited. Identity resolution admitted 160 works carrying 228 claims; 20 works carrying 29 claims remain held, and 0 works carrying 0 claims were excluded.
- Combined public library: 439 admitted works and 567 work claims after canonical-ID augmentation. 8 overlap augmentations add provenance or claims without duplicating a work; 1 merge conflicts remain held out.
- Seed context: 886 identifiers queued for wave 1: 804 identifiers from the research estate plus 82 curated personal-reference identifiers. The seed authority is context, not the publication base.
Partial citation and author coverage
The completed OpenAlex enrichment cache covers the sealed pre-increment snapshot of 179 works. It was not widened to B1/E1 works admitted in this cut: those works contribute no outgoing citation or author-index metadata this round, although an enriched source may cite them as an incoming target. Across the final corpus, 234 of 413 identified works are outside the enrichment snapshot, and 26 admitted works lack a resolved OpenAlex identity—including grey and institutional records. Those unidentified works contribute no citation edges.
The site publishes 793 source-supplied cites edges only where both endpoints are admitted. Across enriched sources, 13,412 referenced-work IDs were observed; 12,619 references to non-corpus works, 0 canonical self-references, and 0 duplicate canonical endpoint pairs were dropped, not held.
The enrichment cache observed 596 authorship occurrences: 592 carried an OpenAlex Author ID and 4 did not. After collapsing 2 exact duplicate A-ID occurrences, the derived index contains 526 A-ID groups and 590 indexed authorships across 178 admitted works. ID-less names remain listed only among the 1,065 bibliographic author-string occurrences on their work records; they are never merged by name and create no author page.
Frozen discovery lanes
Three chartered campaigns were frozen before completion on 2026-08-26 as honest partials. Their evidence records are not in this public corpus:
- Strategy journals (C2): 5 of 12 chartered handoffs delivered.
- Information-systems journals (C3): 4 of 13 delivered.
- Computer-science venues (D1): 3 of 14 delivered; several venue cells never started.
The fractions describe delivery of chartered work packages, not shares of the literature. Planned lanes C4–C6 were never started. Disclosure here does not import frozen-lane evidence content.
Rights and public boundary
The site publishes bibliographic metadata and paraphrased claims only—no source full text, no verbatim evidence quotation, and no vault-relative capture path, regardless of licence or age. Public objects are constructed from fixed field allowlists and scanned again after the static build.
Custody disclosure
On 2026-08-22 a commander-side bulk Git operation may have touched working files in frozen, unpublished lanes. Every release included here is content-verified against immutable release hashes; frozen working files are not publication inputs.