discadmin/disclosure-bureau
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
disclosure-bureau/CLAUDE-schema-full.md
guto 4459bd17e4 phase-0: kill stubs, ship 20 curated anchor events, configure SMTP 
- scripts/03-dedup-entities.py: stop emitting placeholder narrative ("Stub. Will
  be enriched in Phase 7"); write summary_status=none + null fields instead.
- scripts/maintain/41_strip_stubs.py: idempotent migration that cleaned the
  22,096 entity .md files (now zero stub strings in wiki/).
- scripts/synthesize/01_anchor_events.py: curated 20 anchor UAP events
  (Roswell, Nimitz Tic-Tac, Phoenix Lights, Operação Prato, AATIP, etc.) with
  bilingual Holmes-Watson narrative via claude -p --model sonnet
  (CLAUDE_CODE_OAUTH_TOKEN). All summary_status=curated, confidence=high.
- web/api/timeline + timeline-view: filter narrative-less events by default,
  render "curado" badge for hand-vetted ones, drop the date display alone.
- CLAUDE-schema-full.md: document the summary_status enum and the four states.
- docker-compose.yml: SMTP_HOST=mail.spacemail.com configured;
  GOTRUE_MAILER_AUTOCONFIRM flipped to false (real email confirmation working).
- .nirvana/outputs/.../systems-atelier/: 5 deliverables of the architecture
  audit that produced this roadmap.
2026-05-18 00:44:17 -03:00

1220 lines
37 KiB
Markdown
Raw Permalink Blame History

# CLAUDE-schema-full.md — Schema Canônico dos 24 Tipos
> Versão `0.1.0` · Companheiro de [`CLAUDE.md`](CLAUDE.md). Toda página `.md` em `wiki/` e `case/` valida contra um dos 24 tipos abaixo.
## 0.1 Bilingual field convention
For every narrative text field listed below, there is an implicit `<field>_pt_br` sibling holding the **Brazilian Portuguese** translation. The vision call generates both at once. Examples:
| EN field | PT-BR sibling |
|---|---|
| `vision_description` | `vision_description_pt_br` |
| `narrative_summary` (events) | `narrative_summary_pt_br` |
| `executive_summary` (documents — when present as a structured field) | `executive_summary_pt_br` |
| `description` (gaps) | `description_pt_br` |
| `definition_short` (concepts) | `definition_short_pt_br` |
| `verdict_rationale` (witnesses) | `verdict_rationale_pt_br` |
| `connection_description` (relations) | `connection_description_pt_br` |
Fields that are NEVER translated (always in source language):
- OCR text, `verbatim_excerpt`, `verbatim_quotes`, `caption_ocr`
- Enums (`page_type`, `content_classification`, `evidence_grade`, etc.)
- Classification markings (`SECRET//NOFORN`), redaction codes (`(b)(1) 1.4(a)`)
- `canonical_name` (use `aliases[]` for PT-BR forms)
## 0.2 The `war_gov` frontmatter block (injected by 02b-enrich-with-web-metadata.py)
Documents matched against the war.gov metadata JSON get a `war_gov` block injected into the top-level frontmatter. The block is **never overwritten** by re-runs — only created or updated as new fields appear.
```yaml
war_gov:
record_id: record-061 # canonical war.gov id (record-001..record-NNN)
title_official: "DOW-UAP-D054, ..." # title as listed on war.gov (uppercase, comma-separated)
agency_official: "DEPARTMENT OF WAR"
release_date_official: "2026-05-08" # ISO YYYY-MM-DD (parsed from "5/8/26")
release_date_raw: "5/8/26" # verbatim from portal
incident_date_official: "1947-12-30" # ISO; "NA" when source is "N/A" or empty
incident_date_raw: "12/30/47"
incident_date_confidence: high # high | medium | low | none | speculation
incident_location_official: "MEDITERRANEAN SEA"
document_type_official: ".PDF" # .PDF | .VID | .IMG
match_reason: "primary-id=dow-uap-d54" # how the matcher tied this doc to the record
availability: downloaded # downloaded | pending-upstream (placeholder bug on portal)
extracted_from_war_gov_at: "2026-05-13T14:43:22Z"
```
When `incident_date_official ≠ "NA"` and the document still has `document_date: "NA"` at top-level, the enricher promotes the incident_date to fill `document_date`.
Records `record-140`, `record-154`, `record-155`, `record-156` get `availability: pending-upstream` because the war.gov backend serves placeholder files for them (see `memory/project-war-gov-placeholders.md`).
## 0. Common types
```yaml
# ConfidenceBand
confidence_band: high | medium | low | speculation
# EvidenceGrade (Locard)
# A = primary source, autenticada, cadeia limpa
# B = primary source, cadeia com 1 gap menor
# C = secondary source, paraphrase oficial
# D = secondary source, terceira mão
# E = uncorroborated single witness
# F = anecdotal, hearsay
# BBox (normalizado 0..1 nas coordenadas da página)
bbox:
x: 0.123
y: 0.456
w: 0.234
h: 0.089
# ClassificationMarking
classification_markings:
- level: SECRET # UNCLASSIFIED | CUI | CONFIDENTIAL | SECRET | TOP SECRET
caveats: [NOFORN, ORCON]
location: header # header | footer | banner | stamp
bbox: { x: 0.0, y: 0.0, w: 1.0, h: 0.04 }
# Redaction
redactions:
- code: "(b)(1) 1.4(a)"
description: "national defense"
bbox: { x: 0.2, y: 0.45, w: 0.4, h: 0.03 }
text_inferred: null
```
## 1. `document` — `wiki/documents/<doc-id>.md`
```yaml
---
schema_version: "0.1.0"
type: document
doc_id: dow-uap-d54-mission-report-mediterranean-sea-na
canonical_title: "Mission Report — Mediterranean Sea (date NA)"
original_filename: "DOW-UAP-D54-Mission-Report-Mediterranean-Sea-NA.pdf"
raw_path: "../raw/DOW-UAP-D54-Mission-Report-Mediterranean-Sea-NA.pdf"
sha256: "a3f2..."
size_bytes: 20095
page_count: 7
mime_type: "application/pdf"
collection: DOW-UAP # DOW-UAP | DOS-UAP | NASA-UAP | FBI-Vault | FBI-Photo | Incident-Summaries | FOIA-059UAP | Numeric-Files | Other
document_class: mission-report # mission-report | range-fouler-debrief | email-correspondence | diplomatic-cable | apollo-transcript | crew-debriefing | photograph | incident-summary | foia-release | composite-sketch | event-slides | unknown
content_classification:
- text-only
- contains-tables
- redaction-heavy
provenance:
source_url: "https://war.gov/ufo/..."
agency: "Department of War"
release_authority: "DoD AARO"
release_date: "2024-11-15"
foia_request_id: null
document_date: "NA" # YYYY-MM-DD | YYYY-MM | YYYY | "NA"
ingest_date: "2026-05-13"
last_ingest: "2026-05-13T14:22:11Z"
last_lint: "2026-05-13T15:04:00Z"
wiki_version: "0.1.0"
highest_classification: "SECRET//NOFORN"
has_redactions: true
redaction_codes_present: ["(b)(1) 1.4(a)", "(b)(3)"]
languages_detected: ["en"]
executive_summary_confidence: medium
key_claims:
- text: "F/A-18 piloto observa objeto esférico a 25.000 ft no Mediterrâneo"
confidence_band: high
evidence_refs: ["[[evidence/E-0042]]", "[[evidence/E-0043]]"]
page_refs: ["[[dow-uap-d54-.../p003]]", "[[dow-uap-d54-.../p007]]"]
pages:
- page: 1
page_id: "[[dow-uap-d54-.../p001]]"
page_type: cover
classification: "SECRET//NOFORN"
# … até page_count
key_entities:
people: ["[[people/redacted-pilot-01]]"]
organizations: ["[[org/aaro]]", "[[org/uss-gerald-r-ford-cvn-78]]"]
locations: ["[[loc/mediterranean-sea]]"]
events: ["[[event/EV-XXXX-XX-XX-mediterranean-sphere]]"]
uap_objects: ["[[uap/OBJ-EVMED-01]]"]
evidence_extracted: ["[[evidence/E-0042]]"]
witnesses_extracted: ["[[witness/W-0007]]"]
gaps_flagged: ["[[gap/G-0012]]"]
related_documents: ["[[dow-uap-d3-mission-report-arabian-gulf-2020]]"]
external_sources:
- url: "https://media.defense.gov/aaro/uap-historical-record-2024.pdf"
fetched_at: "2026-05-13T14:30:00Z"
summary: "AARO Historical Record cita este caso na p.47"
confidence_band: high
---
```
**Corpo:** `# Title` · `## Sumário Executivo` (com footnotes `[^E-NNNN]`) · `## Índice de Páginas` · `## Entidades-Chave` · `## Conexões com Outros Documentos` · `## Gaps e Anomalias` · `## Procedência` · `## Footnotes`.
## 2. `page` — `wiki/pages/<doc-id>/p<NNN>.md`
```yaml
---
schema_version: "0.1.0"
type: page
page_id: "dow-uap-d54-mission-report-mediterranean-sea-na/p007"
doc_id: "dow-uap-d54-mission-report-mediterranean-sea-na"
page_number: 7
total_pages: 7
png_path: "../../processing/png/dow-uap-d54-.../p007.png"
png_sha256: "9c2e..."
png_dpi: 200
png_width: 1700
png_height: 2200
ocr_raw_path: "../../processing/ocr/dow-uap-d54-.../p007.txt"
vision_raw_path: "../../processing/vision/dow-uap-d54-.../p007.json"
vision_model: "claude-haiku-4-5"
vision_run_at: "2026-05-13T13:45:22Z"
page_type: body # cover | toc | body | signature | photo | sketch | map | stamp | blank | appendix | redaction-heavy | table-page | mixed
content_classification:
- text-only
- contains-tables
language_detected: "en"
classification_markings:
- level: SECRET
caveats: [NOFORN]
location: header
bbox: { x: 0.0, y: 0.0, w: 1.0, h: 0.035 }
- level: SECRET
caveats: [NOFORN]
location: footer
bbox: { x: 0.0, y: 0.965, w: 1.0, h: 0.035 }
redactions:
- code: "(b)(1) 1.4(a)"
description: "national defense"
bbox: { x: 0.18, y: 0.42, w: 0.42, h: 0.025 }
text_inferred: null
- code: "(b)(6)"
description: "personal privacy"
bbox: { x: 0.21, y: 0.51, w: 0.18, h: 0.022 }
text_inferred: "[pilot name]"
signatures_observed:
- signer_inferred: "[[people/redacted-co-pilot-01]]"
confidence_band: low
bbox: { x: 0.55, y: 0.88, w: 0.20, h: 0.04 }
notes: "Assinatura ilegível, anotação datilografada abaixo: 'LCDR'"
tables_detected:
- table_id: "[[table/TBL-DOWD54-0003]]"
bbox: { x: 0.10, y: 0.30, w: 0.80, h: 0.35 }
spans_multi_page: true
continues_from: "[[dow-uap-d54-.../p006]]"
continues_to: null
images_detected:
- image_id: "[[image/IMG-DOWD54-p007-01]]"
image_type: sketch # photo | sketch | map | chart | stamp | signature | redaction | logo | seal | diagram | other
bbox: { x: 0.20, y: 0.68, w: 0.60, h: 0.22 }
caption_ocr: "Fig. 3 — Object trajectory as observed"
entities_extracted:
people: ["[[people/redacted-pilot-01]]"]
organizations: ["[[org/cvw-7]]", "[[org/uss-gerald-r-ford-cvn-78]]"]
locations: ["[[loc/mediterranean-sea]]"]
events: ["[[event/EV-XXXX-XX-XX-mediterranean-sphere]]"]
uap_objects: ["[[uap/OBJ-EVMED-01]]"]
vehicles: ["[[vehicle/fa-18-super-hornet]]"]
operations: []
concepts: ["[[concept/range-fouler]]"]
uap_observation_fields: # presente apenas quando aplicável
date_time_utc: "NA"
duration_seconds: 180
shape: sphere
color: "white-metallic"
size_estimate: "1-3 m"
altitude_ft: 25000
speed_kts: "stationary-then-300"
bearing_deg: 270
distance_nm: 2.5
coordinates:
lat: null
lon: null
confidence_band: low
ocr_quality_score: 0.87
vision_quality_score: 0.92
flags: [] # ["low-ocr", "heavy-redaction", "rotated", ...]
last_ingest: "2026-05-13T13:45:22Z"
last_lint: "2026-05-13T15:04:00Z"
wiki_version: "0.1.0"
---
```
**Corpo:** `# [[doc-id]] — Página N de M` · imagem `![Page N](png_path)` · `## Texto OCR (raw, idioma original)` em blockquote · `## Descrição Vision` (idioma original) · `## Tabelas` · `## Imagens` · `## Entidades nesta página` · `## Observações de Investigação` (curto, PT-BR).
## 3. `person` — `wiki/entities/people/<id>.md`
```yaml
---
schema_version: "0.1.0"
type: entity
entity_class: person
person_id: "david-grusch"
canonical_name: "David Charles Grusch"
aliases: ["David Grusch", "D. Grusch"]
display_name: "David Grusch"
roles:
- title: "Intelligence Officer"
organization: "[[org/nro]]"
period: { start: "2021", end: "2023", confidence_band: medium }
dates:
born: null
died: null
active_period: { start: "~2000", end: "ongoing", confidence_band: low }
mentioned_in: # populado pelo Lint
- page: "[[dow-uap-d54-.../p007]]"
mention_count: 2
role_in_page: subject # subject | witness | author | signer | mentioned
total_mentions: 3
documents_count: 2
signatures_observed:
- page: "[[dow-uap-d54-.../p022]]"
bbox: { x: 0.55, y: 0.88, w: 0.20, h: 0.04 }
confidence_band: medium
verbatim_quotes: # idioma original
- text: "We are not alone in the cosmos."
page: "[[dow-uap-d54-.../p015]]"
bbox: { x: 0.10, y: 0.30, w: 0.80, h: 0.04 }
context: "Congressional testimony, July 2023"
confidence_band: high
related_people:
- person: "[[people/karl-nell]]"
relation_type: colleague # colleague | superior | subordinate | family | source | adversary | unknown
confidence_band: medium
evidence_refs: ["[[evidence/E-0051]]"]
related_organizations: ["[[org/nro]]", "[[org/uaptf]]"]
related_events: ["[[event/EV-2023-07-26-grusch-testimony]]"]
enrichment_status: deep # none | shallow | deep
external_sources:
- url: "https://oversight.house.gov/release/transparency-on-uaps/"
fetched_at: "2026-05-13T14:50:00Z"
summary: "Transcript do testimony de Grusch ao House Oversight."
confidence_band: high
profile_dossier: "[[profile/AP-0001]]"
witness_analyses: ["[[witness/W-0007]]"]
evidence_anchored: ["[[evidence/E-0042]]"]
disambiguation_note: null
last_ingest: "2026-05-13T14:22:11Z"
last_lint: "2026-05-13T15:04:00Z"
wiki_version: "0.1.0"
---
```
**Corpo:** `# Display Name` · `## Identidade` · `## Biografia Interna` (extraída dos PDFs, footnotes) · `## Biografia Externa` (enrichment, footnotes) · `## Quotes Verbatim` · `## Relações` · `## Aparições no Corpus` · `## Procedência Externa`.
## 4. `organization` — `wiki/entities/organizations/<id>.md`
```yaml
---
schema_version: "0.1.0"
type: entity
entity_class: organization
organization_id: "aaro"
canonical_name: "All-domain Anomaly Resolution Office"
aliases: ["AARO"]
parent_organization: "[[org/dod-ousd-i-and-s]]"
child_organizations: []
organization_type: government-agency # government-agency | military-unit | intelligence-agency | corporation | ngo | think-tank | other
country: "USA"
founded: "2022-07-15"
dissolved: null
predecessors: ["[[org/aoimsg]]"]
successors: []
mentioned_in: [...]
total_mentions: 47
documents_count: 23
key_people:
- person: "[[people/jon-kosloski]]"
role: Director
period: { start: "2024", end: ongoing, confidence_band: high }
related_organizations: ["[[org/dod]]", "[[org/odni]]"]
operations_run: ["[[op/aaro-historical-record-report]]"]
enrichment_status: deep
external_sources: [...]
last_ingest: "..."
last_lint: "..."
wiki_version: "0.1.0"
---
```
## 5. `location` — `wiki/entities/locations/<id>.md`
```yaml
---
schema_version: "0.1.0"
type: entity
entity_class: location
location_id: "strait-of-hormuz"
canonical_name: "Strait of Hormuz"
aliases: ["Estreito de Ormuz"]
location_type: strait # city | region | country | sea | strait | airbase | naval-base | mountain | desert | other
country: ["IR", "OM", "AE"]
region: "Persian Gulf"
parent_location: "[[loc/persian-gulf]]"
coordinates:
lat: 26.566667
lon: 56.25
confidence_band: high
source: "Wikipedia, cross-checked"
mentioned_in: [...]
total_mentions: 12
documents_count: 4
events_here: ["[[event/EV-2020-09-XX-strait-of-hormuz-sphere]]"]
related_locations: ["[[loc/persian-gulf]]"]
enrichment_status: shallow
external_sources: [...]
last_ingest: "..."
wiki_version: "0.1.0"
---
```
## 6. `event` — `wiki/entities/events/<id>.md`
```yaml
---
schema_version: "0.1.0"
type: entity
entity_class: event
event_id: "EV-2004-11-14-tic-tac-nimitz"
canonical_name: "Encontro Tic-Tac do Nimitz"
aliases: ["Nimitz Encounter", "Tic Tac Incident"]
event_class: uap-encounter # uap-encounter | testimony | hearing | foia-release | crash-recovery | photograph | declassification | other
date_start: "2004-11-14"
date_end: "2004-11-14"
time_start_utc: "approx 11:00"
duration_estimate_minutes: 180
date_confidence: high
primary_location: "[[loc/pacific-ocean-off-san-diego]]"
locations_involved: ["[[loc/pacific-ocean-off-san-diego]]"]
coordinates: { lat: 31.42, lon: -117.13, confidence_band: high }
observers:
- person: "[[people/david-fravor]]"
role: "primary witness, pilot"
witnesses_analyses: ["[[witness/W-0001]]"]
organizations_involved: ["[[org/uss-nimitz-cvn-68]]"]
vehicles_involved: ["[[vehicle/fa-18-super-hornet]]"]
uap_objects: ["[[uap/OBJ-EV2004-NIMITZ-01]]"]
documented_in: [...]
total_mentions: 18
documents_count: 7
narrative_summary: |
Em 14 de novembro de 2004, durante exercícios do CSG-11...
narrative_summary_pt_br: |
Em 14 de novembro de 2004, durante os exercícios do CSG-11...
# Synthesis pipeline status (Fase 0+3). Replaces the older
# `narrative_summary_confidence` field which conflated provenance and quality.
#
# none — narrative absent. Renderer must show badge "sem síntese ainda"
# and link to chunks via entity_mentions. NEVER emit placeholder
# text like "_Stub. Will be enriched in Phase N._".
# synthesized — produced automatically by scripts/synthesize/20_entity_summary.py
# via Claude Code OAuth (Sonnet). summary_confidence default 'medium'.
# curated — written or revised by hand (or by scripts/synthesize/01_anchor_events.py
# for canonical anchor events). summary_confidence default 'high'.
# red_teamed — curated + reviewed by chief-detective for factual accuracy.
summary_status: curated # none | synthesized | curated | red_teamed
summary_confidence: high # high | medium | low | null
related_events: ["[[event/EV-2015-XX-XX-gimbal]]"]
preceded_by: []
followed_by: ["[[event/EV-2017-12-16-nyt-publication]]"]
evidence_anchored: ["[[evidence/E-0010]]"]
hypotheses_addressing: ["[[hypothesis/H-0001]]"]
gaps_flagged: ["[[gap/G-0003]]"]
enrichment_status: deep
external_sources: [...]
last_ingest: "..."
wiki_version: "0.1.0"
---
```
## 7. `uap_object` — `wiki/entities/uap-objects/<id>.md`
```yaml
---
schema_version: "0.1.0"
type: entity
entity_class: uap_object
uap_object_id: "OBJ-EV2004-NIMITZ-01"
canonical_name: "Tic-Tac Object — Nimitz 2004"
observed_in_event: "[[event/EV-2004-11-14-tic-tac-nimitz]]"
secondary_events: []
shape: elongated-ellipsoid # sphere | disc | triangle | cylinder | cube | elongated-ellipsoid | cigar | irregular | unknown
shape_aliases: ["tic-tac", "pill-shape"]
color: "white-matte"
size_estimate_m: { min: 12, max: 14, confidence_band: medium }
features:
- "no visible exhaust"
- "no visible wings"
altitude_ft: { min: 80000, max: 80000, observed_descend_to_ft: 50, confidence_band: medium }
speed_kts: { min: 0, max: "supersonic-instantaneous", confidence_band: medium }
acceleration_g: { estimate: extreme, confidence_band: low }
maneuver_descriptors:
- instantaneous-direction-change
- hover
- descent-80kft-to-50ft-in-seconds
sensor_observations:
- sensor: "AN/SPY-1 radar (USS Princeton)"
type: radar
- sensor: "ATFLIR (F/A-18)"
type: infrared
- sensor: "Mark I eyeball"
type: visual
observers: ["[[people/david-fravor]]"]
visual_records: ["[[image/IMG-DOC65-p014-01]]"]
documented_in: [...]
total_mentions: 12
evidence_anchored: ["[[evidence/E-0010]]"]
hypotheses_addressing: ["[[hypothesis/H-0001]]"]
confidence_band_overall: medium
last_ingest: "..."
wiki_version: "0.1.0"
---
```
## 8. `vehicle` — `wiki/entities/vehicles/<id>.md`
```yaml
---
schema_version: "0.1.0"
type: entity
entity_class: vehicle
vehicle_id: "fa-18-super-hornet"
canonical_name: "F/A-18 Super Hornet"
aliases: ["F/A-18E", "F/A-18F", "Super Hornet"]
vehicle_class: aircraft # aircraft | ship | submarine | spacecraft | satellite | ground | other
manufacturer: Boeing
operator: "[[org/us-navy]]"
service_period: { start: "1995", end: ongoing }
mentioned_in: [...]
total_mentions: 38
events_involved: ["[[event/EV-2004-11-14-tic-tac-nimitz]]"]
sensors: ["ATFLIR", "AN/APG-79 AESA radar"]
enrichment_status: shallow
external_sources: [...]
last_ingest: "..."
wiki_version: "0.1.0"
---
```
## 9. `operation` — `wiki/entities/operations/<id>.md`
```yaml
---
schema_version: "0.1.0"
type: entity
entity_class: operation
operation_id: "range-fouler"
canonical_name: "Range Fouler Program"
aliases: ["Range Fouler"]
operation_type: reporting-protocol # military-operation | reporting-protocol | research-program | task-force | foia-disclosure | other
status: active
period: { start: "~2019", end: ongoing, confidence_band: medium }
run_by: ["[[org/us-navy]]", "[[org/aaro]]"]
key_people: []
description_summary: |
Termo usado pela Marinha dos EUA...
documents: ["[[dow-uap-d38-...]]", "[[dow-uap-d42-...]]"]
total_mentions: 22
related_concepts: ["[[concept/uap-doctrine]]"]
related_events: ["[[event/EV-2019-04-XX-east-coast-incursions]]"]
enrichment_status: shallow
last_ingest: "..."
wiki_version: "0.1.0"
---
```
## 10. `concept` — `wiki/entities/concepts/<id>.md`
```yaml
---
schema_version: "0.1.0"
type: entity
entity_class: concept
concept_id: "foia-exemption-1-4-a"
canonical_name: "FOIA Exemption (b)(1) 1.4(a)"
aliases: ["1.4(a)", "EO 13526 1.4(a)"]
concept_class: legal-instrument # legal-instrument | phenomenon-type | doctrine | scientific-term | jargon | program-name | other
domain: "FOIA / Executive Order 13526"
definition_short: |
Categoria de classificação 'Military plans, weapons systems, or operations'
sob Executive Order 13526 Section 1.4(a)...
mentioned_in: [...]
total_mentions: 84
documents_count: 41
related_concepts: ["[[concept/foia-exemption-b-3]]"]
related_organizations: ["[[org/national-archives]]"]
enrichment_status: deep
external_sources: [...]
last_ingest: "..."
wiki_version: "0.1.0"
---
```
## 11. `table` — `wiki/tables/<table-id>.md`
```yaml
---
schema_version: "0.1.0"
type: table
table_id: "TBL-DOWD54-0003"
canonical_title: "GENTEXT UAP Observation Fields — Mediterranean Sea"
source_doc: "[[dow-uap-d54-...]]"
spans_pages:
- page: "[[dow-uap-d54-.../p006]]"
bbox: { x: 0.10, y: 0.55, w: 0.80, h: 0.40 }
role: start
- page: "[[dow-uap-d54-.../p007]]"
bbox: { x: 0.10, y: 0.30, w: 0.80, h: 0.35 }
role: middle
- page: "[[dow-uap-d54-.../p008]]"
bbox: { x: 0.10, y: 0.10, w: 0.80, h: 0.20 }
role: end
total_rows: 24
total_columns: 4
has_headers: true
multi_page: true
columns:
- { name: "Field", type: string }
- { name: "Value", type: string }
- { name: "Unit", type: string }
- { name: "Confidence", type: string }
row_count_extracted: 24
extraction_quality: 0.91
references:
events: ["[[event/EV-XXXX-XX-XX-mediterranean-sphere]]"]
uap_objects: ["[[uap/OBJ-EVMED-01]]"]
last_ingest: "..."
wiki_version: "0.1.0"
---
```
**Corpo:** tabela reconstruída em markdown nativo + bloco "Notas sobre reconstrução" (cells perdidas, ambiguidades).
## 12. `image` — `wiki/images/<image-id>.md`
```yaml
---
schema_version: "0.1.0"
type: image
image_id: "IMG-DOWD54-p007-01"
image_type: sketch # photo | sketch | map | chart | stamp | signature | redaction | logo | seal | diagram | other
source_page: "[[dow-uap-d54-.../p007]]"
bbox_on_page: { x: 0.20, y: 0.68, w: 0.60, h: 0.22 }
extracted_png_path: "../../processing/images/IMG-DOWD54-p007-01.png"
extracted_sha256: "7a1f..."
caption_ocr: "Fig. 3 — Object trajectory as observed" # idioma original
caption_inferred: null # PT-BR opcional, futuro
vision_description: |
Sketch a lápis em folha pautada. Mostra silhueta de F/A-18 à esquerda
em perfil, com seta tracejada saindo de altitude alta e descendo em
zig-zag até 'sea level'. Objeto esférico representado como círculo
hachurado. Anotações manuscritas: 'observed 270° rel', 'descent ~5s'.
vision_quality_score: 0.94
ocr_quality_score: 0.62
contains:
vehicles: ["[[vehicle/fa-18-super-hornet]]"]
uap_objects: ["[[uap/OBJ-EVMED-01]]"]
events: ["[[event/EV-XXXX-XX-XX-mediterranean-sphere]]"]
forensic_metadata:
exif_present: false
origin_camera: null
date_taken: null
geolocation: null
manipulation_detected: null # null | none | crop | color-shift | suspected
evidence_refs: ["[[evidence/E-0042]]"]
related_images: ["[[image/IMG-DOWD54-p012-01]]"]
confidence_band_overall: medium
last_ingest: "..."
wiki_version: "0.1.0"
---
```
## 13. `evidence` — `case/evidence/<E-NNNN>.md`
```yaml
---
schema_version: "0.1.0"
type: evidence
evidence_id: "E-0042"
canonical_title: "F/A-18 piloto observa esfera branco-metálica a 25.000 ft no Mediterrâneo"
evidence_grade: B # A | B | C | D | E | F
evidence_class: testimonial-primary # physical | testimonial-primary | testimonial-secondary | documentary-primary | documentary-secondary | photographic | sensor-data | sketch | inferential
source_page: "[[dow-uap-d54-.../p007]]"
source_doc: "[[dow-uap-d54-...]]"
source_bbox: { x: 0.10, y: 0.30, w: 0.80, h: 0.18 }
verbatim_excerpt: |
"...observed a single, stationary, white object approximately 1-3 meters
in diameter at angels 25. Object remained stationary for approximately
180 seconds before departing on bearing 270 at estimated 300 knots…"
chain_of_custody:
- step: 1
action: "Original observation"
actor: "[[people/redacted-pilot-01]]"
location: "[[loc/mediterranean-sea]]"
timestamp: "NA"
confidence_band: medium
- step: 2
action: "Debrief recorded"
actor: "[[org/cvw-7-intelligence]]"
timestamp: "NA"
confidence_band: medium
- step: 3
action: "Document released via war.gov"
actor: "[[org/dod]]"
timestamp: "2024-11-15"
confidence_band: high
custody_gaps:
- between_steps: [1, 2]
gap_description: "Intervalo entre observação e debrief desconhecido."
severity: minor
corroborating_evidence: ["[[evidence/E-0043]]"]
contradicting_evidence: []
supports_claims:
- claim: "Objeto esférico branco-metálico observado no Mediterrâneo"
confidence_band: high
- claim: "Velocidade observada: estacionário → 300 kts"
confidence_band: medium
relevant_to_hypotheses: ["[[hypothesis/H-0001]]"]
locard_principle_applied: [transfer, individuality]
catalogued_by: evidence-officer
catalogued_at: "2026-05-13T14:30:00Z"
reviewed_by: chief-detective
reviewed_at: "2026-05-13T15:00:00Z"
wiki_version: "0.1.0"
---
```
## 14. `witness_analysis` — `case/witnesses/<W-NNNN>.md`
```yaml
---
schema_version: "0.1.0"
type: witness_analysis
witness_id: "W-0007"
witness_person: "[[people/david-fravor]]"
event_witnessed: "[[event/EV-2004-11-14-tic-tac-nimitz]]"
statements:
- statement_id: "S-0007-01"
source_page: "[[doc-.../p015]]"
bbox: { x: 0.10, y: 0.20, w: 0.80, h: 0.40 }
verbatim: "I have no idea what I saw…"
date_given: "2017-12-16"
venue: "NYT interview"
confidence_band_authenticity: high
paraphrase: |
Fravor descreve aproximação ao objeto, manobras em "L" e desaparecimento
do radar...
corroboration:
- corroborator: "[[witness/W-0008]]"
overlap_pct: 78
divergences:
- "Dietrich estima altitude inicial 20kft; Fravor estima 80kft"
severity: minor
contradictions: []
lexical_shifts:
- between_statements: ["S-0007-01", "S-0007-02"]
shift_type: certainty # certainty | tense | emotional-valence | technical-vocabulary | role-attribution
description: "Em 2017 disse 'no idea'; em 2023 caracterizou como 'tecnologia'."
confidence_band: medium
verdict: credible-with-caveats # credible | credible-with-caveats | inconclusive | inconsistent | likely-fabrication
verdict_rationale: |
Testemunho técnico, consistente em núcleo factual ao longo de 6 anos...
reviewed_by: witness-officer
reviewed_at: "2026-05-13T14:40:00Z"
quality_gate_score: 0.89
wiki_version: "0.1.0"
---
```
## 15. `timeline` — `case/timelines/<scope>.md`
```yaml
---
schema_version: "0.1.0"
type: timeline
timeline_scope: global # global | event-cluster | region | actor | decade
scope_id: global
canonical_title: "Timeline Mestre — UFO/UAP Corpus war.gov 1940s-2025"
period: { start: "1940-01-01", end: "2025-12-31" }
entries_count: 187
entries:
- timestamp: "1947-07-XX"
timestamp_confidence: medium
event: "[[event/EV-1947-07-XX-roswell-incident]]"
location: "[[loc/roswell-new-mexico-usa]]"
actors: ["[[org/usaaf]]"]
summary: "Recuperação de destroços em rancho perto de Roswell."
evidence_refs: ["[[evidence/E-0001]]"]
confidence_band: high
clusters:
- cluster_id: C-001
label: "1947 Wave"
period: { start: "1947-06", end: "1947-09" }
entries_count: 14
reconstructed_by: timeline-analyst
reconstructed_at: "..."
wiki_version: "0.1.0"
---
```
## 16. `hypothesis` — `case/hypotheses/<H-NNNN>.md`
```yaml
---
schema_version: "0.1.0"
type: hypothesis
hypothesis_id: "H-0001"
canonical_title: "Tic-Tac é tecnologia humana avançada sob teste"
scope_event: "[[event/EV-2004-11-14-tic-tac-nimitz]]"
hypothesis_class: prosaic-advanced-tech # prosaic-mundane | prosaic-advanced-tech | misidentification | sensor-artifact | psyop | hoax | extraterrestrial | extradimensional | unknown
status: active # active | eliminated | confirmed-best-explanation | dormant
elimination_reason: null
prior_probability: 0.35
posterior_probability: 0.18
posterior_updated_at: "..."
posterior_method: "Tetlock-style structured judgment, 3 detectives"
falsification_tests:
- test_id: "FT-0001-01"
description: "Se for tech US, deve haver registro orçamentário em SAP."
status: untestable-current-data
- test_id: "FT-0001-02"
description: "Se for tech adversária, deve haver capacidade em outros teatros."
status: weakly-falsified
evidence_for: ["[[evidence/E-0015]]"]
evidence_against: ["[[evidence/E-0010]]"]
competes_with: ["[[hypothesis/H-0002]]"]
elimination_logic: |
"Quando você elimina o impossível..." — Esta hipótese sobrevive ao teste...
steel_man: |
Defesa mais forte: programa black SAP US...
red_team: |
Ataque mais forte: SAP US não testaria contra ativos próprios...
evaluated_by: hypothesis-lead
last_reviewed: "..."
quality_gate_score: 0.87
wiki_version: "0.1.0"
---
```
## 17. `actor_profile` — `case/profiles/<AP-NNNN>.md`
```yaml
---
schema_version: "0.1.0"
type: actor_profile
actor_profile_id: "AP-0001"
actor: "[[people/david-grusch]]"
actor_type: person # person | organization
motive:
description: "Whistleblower motivado por convicção de cover-up sistêmico."
evidence_refs: ["[[evidence/E-0050]]"]
confidence_band: medium
means:
description: "Acesso a NRO e UAPTF; clearance TS/SCI."
evidence_refs: ["[[evidence/E-0051]]"]
confidence_band: high
opportunity:
description: "Janela 2019-2023 com acesso direto a programas relevantes."
confidence_band: high
modus_operandi:
description: "Disclosure via canal oficial IC IG, seguido de testimony público."
patterns:
- "Procedural-first disclosure"
confidence_band: medium
baseline: "Intelligence officer típico mantém silêncio pós-clearance."
deviation_signal: "Quebra significativa: testimony público sob juramento."
deviation_severity: high
connections:
- actor: "[[people/karl-nell]]"
nature: professional-corroborator
confidence_band: medium
profiled_by: profiler
profiled_at: "..."
quality_gate_score: 0.88
wiki_version: "0.1.0"
---
```
## 18. `gap` — `case/gaps/<G-NNNN>.md`
```yaml
---
schema_version: "0.1.0"
type: gap
gap_id: "G-0012"
canonical_title: "Ausência de timestamp em DOW-UAP-D54"
gap_class: missing-data # missing-data | inconsistency | unexplained-redaction | chronology-conflict | actor-not-identified | sensor-mismatch | other
description: |
Documento DOW-UAP-D54 referencia 'date NA' no título e omite timestamp...
detected_in: ["[[dow-uap-d54-.../p001]]"]
detected_by: archivist
detected_at: "..."
severity: medium # low | medium | high | critical
investigative_impact: |
Sem timestamp, correlação com sensor data de outras plataformas fica impedida.
possible_explanations:
- { explanation: "Redaction pré-release não documentada", confidence_band: medium }
- { explanation: "Erro de processing OCR", confidence_band: low }
recommended_actions:
- "Cross-check com FOIA request log war.gov"
related_gaps: ["[[gap/G-0008]]"]
wiki_version: "0.1.0"
---
```
## 19. `relation` (connect-the-dots) — `case/connect-the-dots/<R-NNNN>.md`
```yaml
---
schema_version: "0.1.0"
type: relation
relation_id: "R-0028"
canonical_title: "DOW-UAP-D54 e D55 descrevem o mesmo cluster Mediterrâneo"
relation_class: documentary-overlap # documentary-overlap | actor-bridge | location-cluster | sensor-corroboration | temporal-sequence | other
nodes:
- "[[dow-uap-d54-...]]"
- "[[dow-uap-d55-...]]"
- "[[event/EV-XXXX-XX-XX-mediterranean-sphere]]"
connection_description: |
D54 (date NA) e D55 (Syria, Nov 2016) ambos referenciam observações de
objeto esférico branco-metálico, mesma assinatura visual...
connection_strength: 0.72
strength_method: "Jaccard sobre descritores UAP + co-location ± 2000 km"
confidence_band: medium
supporting_evidence: ["[[evidence/E-0042]]", "[[evidence/E-0048]]"]
illuminates_hypotheses: ["[[hypothesis/H-0007]]"]
drawn_by: chief-detective
drawn_at: "..."
wiki_version: "0.1.0"
---
```
## 20. `case_report` — `case/case-report.md`
```yaml
---
schema_version: "0.1.0"
type: case_report
case_id: "CASE-0001"
canonical_title: "O Departamento da Guerra — Memorando ao Dr. Watson sobre 129 Documentos UAP"
narrative_style: holmes-watson
narrator_persona: "Dr. John H. Watson"
detective_persona: "Sherlock Holmes (composite: Holmes/Poirot/Dupin)"
investigation_period: { start: "2026-05-13", end: null }
documents_analyzed: 129
pages_analyzed: null
entities_catalogued: null
evidence_catalogued: null
hypotheses_evaluated: null
executive_finding: |
Após exame de cento e vinte e nove documentos, este investigador conclui...
executive_finding_confidence: medium
chapters:
- chapter: "I — O Recebimento dos Autos"
summary: "..."
- chapter: "II — A Catalogação Forense (Locard)"
summary: "..."
- chapter: "III — As Testemunhas"
summary: "..."
- chapter: "IV — As Linhas do Tempo"
summary: "..."
- chapter: "V — O Tribunal de Hipóteses"
summary: "..."
- chapter: "VI — Conexões Improváveis"
summary: "..."
- chapter: "VII — Lacunas Inegáveis"
summary: "..."
- chapter: "VIII — Conclusão e Incerteza Residual"
summary: "..."
top_evidence: ["[[evidence/E-0010]]"]
top_hypotheses_surviving: ["[[hypothesis/H-0001]]"]
top_hypotheses_eliminated: ["[[hypothesis/H-0004]]"]
critical_gaps: ["[[gap/G-0012]]"]
residual_uncertainty_ref: "[[case/residual-uncertainty]]"
quality_rubrics:
evidence_chain_complete: true
hypothesis_tournament_diverse: true
contradictions_addressed: true
confidence_calibrated: true
procedure_documented: true
falsifiability_explicit: true
overall_quality_score: 0.91
written_by: case-writer
reviewed_by: [chief-detective, hypothesis-lead, evidence-officer]
last_revised: "..."
wiki_version: "0.1.0"
---
```
**Corpo:** narrativa Holmes-Watson em **PT-BR**, prosa elaborada, com inline `[^E-NNNN]`, `[^H-NNNN]`, `[^G-NNNN]` apontando para os artefatos. Cada chapter um `##`. **Quotes verbatim no idioma original** (sem tradução).
## 21. `residual_uncertainty` — `case/residual-uncertainty.md`
```yaml
---
schema_version: "0.1.0"
type: residual_uncertainty
linked_case: "[[case/case-report]]"
unknowns_known:
- unknown: "Identidade dos pilotos redacted em DOW-UAP-D54"
mitigations: ["Cross-ref com squadron rosters públicos"]
confidence_band: low
unknowns_unknown_disclaimer: |
Existem domínios deste corpus não tocados por nenhum dos 8 detetives...
calibration_table:
- claim: "Tic-Tac é tecnologia humana"
probability: 0.18
confidence_band: low
what_would_change_conclusion:
- "Documento desclassificado com link direto entre Range Fouler e SAP US"
black_swan_scenarios:
- scenario: "Disclosure massivo em 2026-2027 reordena o corpus."
p_in_36_months: 0.20
impact: extreme
written_by: chief-detective
last_revised: "..."
wiki_version: "0.1.0"
---
```
## 22. `index` — `wiki/index.md`
```yaml
---
schema_version: "0.1.0"
type: index
canonical_title: "Wiki UFO/UAP — Departamento da Guerra (war.gov/ufo)"
generated_at: "..."
wiki_version: "0.1.0"
stats:
documents: 129
pages: null
entities:
people: null
organizations: null
locations: null
events: null
uap_objects: null
vehicles: null
operations: null
concepts: null
tables: null
images: null
evidence: null
witnesses: null
hypotheses: null
gaps: null
hubs:
- { label: "Caso central — Holmes Report", target: "[[case/case-report]]" }
- { label: "Timeline mestre", target: "[[case/timelines/global]]" }
- { label: "Hipóteses ativas", target: "case/hypotheses/" }
- { label: "Documentos por coleção", target: "#por-colecao" }
- { label: "Eventos canônicos", target: "#eventos-canonicos" }
---
```
## 23. `log` — `wiki/log.md`
```yaml
---
schema_version: "0.1.0"
type: log
canonical_title: "Append-Only Log — Ingest / Query / Lint"
wiki_version: "0.1.0"
---
```
**Corpo (append-only):**
```
## 2026-05-13T13:45:22Z — INGEST
- operator: archivist
- pdf: DOW-UAP-D54-Mission-Report-Mediterranean-Sea-NA.pdf
- doc_id: dow-uap-d54-mission-report-mediterranean-sea-na
- pages: 7
- vision_model: claude-haiku-4-5
- duration_seconds: 412
- new_entities: { people: 1, organizations: 2, locations: 1, events: 1, uap_objects: 1 }
- new_evidence: 2
- warnings: ["page 12 OCR quality 0.61"]
## 2026-05-13T15:04:00Z — LINT
- operator: archivist
- scope: full
- broken_links: 0
- orphan_entities: 3
- duplicate_canonical_names: 0
- missing_required_fields: 0
- backlinks_rebuilt: 1247
- duration_seconds: 38
```
## 24. Validação — must-have vs nice-to-have
### Universal (todos os tipos)
- `schema_version` (lint bloqueia se ausente)
- `type` (lint bloqueia)
- `canonical_title` OU `canonical_name`
- `wiki_version`
- `last_ingest` OU `last_revised` (warning, não bloqueia)
### Regras estruturais (lint adicional)
1. Todo `[[link]]` resolve. Broken link → **bloqueia**.
2. `entity.mentioned_in` ↔ `page.entities_extracted` consistente. **Bloqueia**.
3. Dois `person_id` distintos com mesmo `canonical_name` sem `disambiguation_note` → **bloqueia**.
4. Evidence grade A → ≥3 custody steps; B → ≥2; C → ≥1.
5. Hypothesis com `posterior_probability > 0.50` → ≥2 `evidence_for`.
6. Para cada `document`, `pages[]` contínuo `1..page_count`.
### Owners (Investigation Bureau)
| Agente | Owners |
|---|---|
| archivist | doc, page, index, log, lint, naming, location, vehicle, operation, concept, table |
| evidence-officer | evidence, image forensic_metadata, uap_object |
| witness-officer | witness_analysis, verbatim_quotes em person |
| timeline-analyst | timeline, event |
| profiler | actor_profile, person, organization |
| hypothesis-lead | hypothesis, falsification_tests |
| chief-detective | relation, gap escalation, residual_uncertainty |
| case-writer | case-report |
Reference in a new issue View git blame Copy permalink