§2As of 2026-07-17

Document audit, provenance chain, and the Stage 1 / Stage 2 recovery

Editorial note: This section is drawn from Phase 8 Final.md. A later document, Final Write Up.md, claimed the Phase 1/6/7 base dossiers had been recovered; an 18 July 2026 checksum audit found this claim to be incorrect — those files were byte-identical duplicates of the correction addenda. Phase 8 Final.md's original framing below is the version confirmed correct by that audit.

§2 — Document audit, provenance chain, and the Stage 1 / Stage 2 recovery

This section closes Phase 0 Tasks 1 and 2, which Phase 0 could not perform.

2.1 Files present and read

#File (as named)True formatExtentRead?
1Project_Handoff_Argentina_Messi_FIFA_CONMEBOL_and_AFA_Documentary.pdfzip of 17 page-images + OCR text17 pp / ~2,900 wordsFully read
2Argentina_Qatar_2022_Officiating_Review.pdfzip, 11 pp~3,900 wordsFully read
3Stage_1_Evidence_Audit_Supplement_on_Argentina_Officiating_and_Discipline.pdfzip, 11 pp~3,500 wordsFully read
4Corrections_Addendum_to_Stage_1.pdfzip, 6 pp~2,300 wordsFully read
5Stage_Two_Report_on_Inter_Miami_and_the_United_States_Host_Berth_for_the_2025_FIFA_Club_World_Cup.pdfzip, 9 pp~3,500 wordsFully read
6Stage_2_Documentary_Record_Addendum.pdfzip, 12 pp~4,200 wordsFully read
7Deep_Research_Framework_for_an_Unspecified_Topic.pdfzip, 8 pp~3,000 wordsFully read — NOT RELATED TO THIS PROJECT
8–15phase-0-setup-schema-and-coverage-map.md, Phase_1.mdPhase_7.mdmarkdown~54,000 wordsFully read

Verification method: file reported all seven .pdf files as "Zip archive data"; pdfinfo/pdffonts failed with "Couldn't find trailer dictionary"; unzip -l revealed N.jpeg + N.txt + manifest.json per document. Text was extracted from the .txt layers in page order. No file was unreadable. No file was present-but-not-attempted.

File 7 is a null. Deep_Research_Framework_for_an_Unspecified_Topic.pdf is a generic research memo about generative AI's effect on enterprise productivity and labor markets. It has no connection to Messi, football, or any project subject. It is recorded here as present-and-irrelevant so no future phase spends effort on it.

2.2 The Stage 1 / Stage 2 restatement Phase 0 could not give

The Project Handoff (11 July 2026) is the operative brief. Stage 1's own report says so: "The original request did not specify a topic, but the uploaded project handoff does." It commissions four Stages in one chat: Stage 1 — Argentina's five Qatar 2022 penalties + the complete Argentina–Netherlands officiating timeline; Stage 2 — the Inter Miami U.S. host berth for the 2025 Club World Cup; Stage 3 — AFA, TourProdEnter and U.S. financial scrutiny; Stage 4 — Messi 2016, Messi 2019, Lahoz–Negreira, and final synthesis. It sets seven claim labels, a twelve-tier source hierarchy, and a hard rule: "Do not use Wikipedia as a final source."

Stage 1 findings (three documents). Argentina were awarded five in-play penalties (Saudi Arabia, Poland, Netherlands, Croatia, France). They fall into different evidentiary tiers: Netherlands and Croatia easiest to defend; Saudi Arabia supportable (holding in the area is a direct-free-kick offence); Poland and France softest and most contestable, but each with a plausible rules basis. The quarter-final set a World Cup record of 18 yellows + 1 red. Stage 1's controlling conclusion: "chaos and poor match control are not the same thing as pro-Argentina bias" and "dysfunction is a better fit than conspiracy." The Corrections Addendum adds that the review process was not uniform — Saudi Arabia and Poland best supported as VAR-initiated with an on-field review; Netherlands, Croatia and France as live awards not shown to have gone through a public OFR.

Stage 2 findings (two documents). Inter Miami earned the 2024 Supporters' Shield on the field; FIFA selected Inter Miami as U.S. host representative on 19 October 2024. The public FIFA materials reviewed did not state the Supporters' Shield criterion before Inter Miami clinched it. The Stage 2 Addendum establishes the documentary spine: host berth created 14 Feb 2023 ("tournament host: 1"); access principles 14 Mar 2023; US appointed host 23 Jun 2023; 17 Dec 2023 Council release "further details being provided in due course"; Garber 24 Jul 2024 says criteria "Not yet" finalized, MLS had "made suggestions to FIFA", FIFA's website said "TBD"; Shield clinched 2 Oct 2024; Infantino names Miami 19 Oct 2024; final regulations published 5 Nov 2024 still say the host slot's access "will be determined by FIFA at a later stage." No public Council vote naming Inter Miami was retrieved. Conclusion: "On the reviewed public record, FIFA's explanation followed the Shield clinch; this review located no pre-clinch published binding criterion."

2.3 Provenance chain, corrected

Original ~25-section brief  ──────► STILL MISSING (never located; not in the Project)
        │
        ├──► Project Handoff (11 Jul 2026, 17pp) ──► Stage 1 report ──► Stage 1 Supplement ──► Corrections Addendum to Stage 1
        │                                        └──► Stage 2 report ──► Stage 2 Documentary Record Addendum
        │                                        └──► Stage 3 ──► NEVER PRODUCED
        │                                        └──► Stage 4 ──► NEVER PRODUCED AS SUCH
        │
        └──► Prior chat "Task assessment and model selection" ──► 9-phase prompt pack ──► Phases 0–8

The two branches never met. The Stage branch and the Phase branch are parallel investigations of the same subject, run from the same missing brief, that did not read each other. Phases 3, 4 and 7 re-researched Stage 4 and Stage 2 material from scratch. Phase 6 re-researched Stage 1 material from scratch. Annexes B and C record what that cost and what it bought.

2.4 Source-quality alert: the Corrections Addendum to Stage 1

The Corrections Addendum to Stage 1 is sourced overwhelmingly from Wikipedia. Its bibliography lists twelve URLs; eleven are Wikipedia (en/es/pt), the twelfth is a FIFA training-centre PDF. This violates the Handoff's own rule ("Do not use Wikipedia as a final source") and Phase 0 §3.2's Wikipedia rule ("It never appears in Original source, Independent confirmation, Direct URL, or the primary-source ledger, and it is never evidence").

Practical effect: the Corrections Addendum's headline "corrected fact sheet" — the two technical-area cautions to Walter Samuel and Lionel Scaloni, and the Cody Gakpo caution — rests on Wikipedia. Phase 6 independently confirms Samuel and Scaloni from the FIFA Full-Time Match Report, so that finding survives on better evidence. Phase 6 does not contain Gakpo at all. See Annex B, C-01.

2.5 Correction-ledger rows whose target entry was never imported

Every row of every phase correction ledger is preserved in this archive. Three of the twenty name an entry ID that has no corresponding record in the imported dataset, for reasons that are themselves a consequence of the checksum audit in the editorial note above. None of the three was resolved by inventing a target entry; each is recorded here so the gap is visible rather than silent.

Ledger rowEntry ID as writtenStatusBasis
Phase 6, row 7TOURN-035, TOURN-036, othersNo live target. Documented gap.Both IDs appear in the corpus only inside the Phase 6 addendum itself, as bare "AMENDED — null-language correction only" stubs carrying an ID, the amendment sentence and a verification date — no event date, headline, summary, classification or source. The base TOURN dossier that would supply that text is not in the corpus (the file supplied as Phase_6_BASE_TOURN.md is byte-identical to the addendum). Neither entry was importable. The row's own third token, "others", is non-exhaustive by construction.
Phase 6, row 9New TOURN-038, TOURN-039Mapped to real records.TOURN-038 is a parent grouping ID that was never a standalone entry even in the addendum, which defines only the incident children TOURN-038-INC01INC05. The row is mapped to the four incidents actually imported plus TOURN-039. TOURN-038-INC03 is deliberately absent: it is a pure cross-reference ("See TOURN-032-INC01") to the Netherlands penalty already recorded under TOURN-032, and importing it would have duplicated one incident under two IDs.
Phase 7, row C6MIA-005Mapped to its successor.MIA-005 is the superseded base entry; the Phase 7 addendum contains no MIA-005 record (the base MIA dossier is likewise not in the corpus). Its corrected successor is MIA-025, on the addendum's own attribution: MIA-025's Relevance field reads "Corrects the 'inaugural' error (C6) and completes the 2023 chronology."

The amendment recorded by Phase 6 row 7 — replacing absolute nulls such as "NONE located" or "no sanction exists" with "no publicly available record was located this session" — is a house wording rule that the imported corpus already follows throughout, independently of whether the two stub entries exist. What is lost with TOURN-035 and TOURN-036 is the entries themselves, not the correction's effect.