OIC-GC-005 — Procurement preflight corpus selection and provenance

[veraxis-protocol]

Owner

Arkadiy Miteiko / Veraxis for corpus authorization and scope; Codex may perform bounded research, acquisition, inventory, and provenance preparation.

Required inputs

• Bounded public or synthetic procurement use-case criteria.
• Source candidates, access terms, provenance metadata, hashes, and handling restrictions.
• Existing preflight SOURCE_MANIFEST.csv contract and Lab restrictions.

Required outputs

• Proposed bounded corpus inventory.
• Completed rights/provenance records for selected candidates.
• Acquisition and hashing evidence, exclusions, uncertainty log, and reviewer-ready source manifest proposal.

Acceptance criteria

• Every candidate has source, jurisdiction, date/version, acquisition method, digest, rights basis, and redistribution/use constraints recorded.
• Confidential, privileged, regulated, export-controlled, and classified content is excluded.
• Unclear rights or provenance remains blocked rather than inferred.
• Selection is reviewed before any extraction work.

Explicit non-goals

• No Docling integration, LLM extraction, normative-unit extraction, admission, IR construction, or semantic interpretation.
• No claim that the corpus is representative or benchmark-superior.

Dependencies

• Licensing questions coordinate with OIC-GC-008.
• Feeds OIC-GC-007.

Current gate impact

Preflight provenance is currently OPEN. Completion supplies owner-review evidence but does not itself open the semantic gate.

[veraxis-protocol]

RECONCILIATION STATUS — 2026-08-17

Historical state (preserved, not overwritten):

• PR research: compare procurement preflight corpus candidates #15 merged (8a992d5fbe536dc35cb5d0cb13e01c447ab09ce7) — procurement preflight corpus candidates compared; Canada selected as the then-bounded source family.
• PR corpus/canada: rights-resolution dossier for the ten blocked sources (v0.1) #21 merged (29daa374b7e5cdc30ca7788310fbabb85f19912b) — Canada rights-resolution dossier. Its ceiling stands: all ten unresolved sources remain BLOCKED_PENDING_CLEARANCE, no source retrieval was performed, no publisher permission was received, no counsel clearance was established.
• PR research: add Morocco official-source preflight #22 open as a draft — Morocco official-source preflight, PROVISIONAL_UNRECONCILED, institutional admission NONE, golden mission NOT_ADMITTED, semantic implementation gate BLOCKED. Untouched by this reconciliation.

What changed in framing:

• The original issue can read as though choosing a corpus family were the objective. It is not. Selecting a family and holding an admissible mission corpus are different states, and PR research: compare procurement preflight corpus candidates #15 established only the former.
• Restated objective: bounded mission corpus provenance and admission readiness.

The stages, kept separate — none implies the next:

1. source discovery
2. source acquisition
3. provenance
4. rights
5. source fidelity
6. normative applicability
7. mission selection
8. institutional admission

PR #15 reached (1) and part of (7) for Canada. PR #21 established that (4) is unresolved for all ten Canada sources. PR #22 is a provisional, unreconciled pass at (1)–(3) for Morocco. No stage beyond that is established for any source.

Current status:

• OPEN — REBASELINED. Not closable on PR research: compare procurement preflight corpus candidates #15; not closable on PR research: add Morocco official-source preflight #22.

Remaining closure conditions:

• Per mission-selected source: retrieval provenance, fidelity to the official publication, rights basis for use and for redistribution, and normative applicability to the mission.
• Rights clearance for anything still BLOCKED_PENDING_CLEARANCE, which today is all ten Canada sources.
• For the current World Bank / Morocco path specifically: the official-source identity and retrieval record, the publication authority, redistribution basis, language-of-record handling, and the applicability determination for the intended mission — recorded as mission evidence, not as a submission narrative.
• An explicit admission decision, by the owner, for a named bounded corpus.

Claim ceiling:

• No source is admitted merely because it is official-domain material. Official publication is a fact about origin, not a grant of rights, not evidence of fidelity, and not an admission decision.
• Nothing here establishes corpus admission, mission readiness, publisher permission, counsel clearance, or any CDC claim.