# Agent bootstrap topology

- **Topology:** `ABT-2026.09.05.3`
- **Living Source:** `YLS-2026.09.05.3`
- **Claims:** 69
- **Approved wisdoms:** 364
- **Teacher journey:** 39 nodes / 76 edges / 11 first-release nodes
- **Identity Mind records:** 1,194
- **Candidate identity-cell shells:** 761
- **Nonsemantic identity-cell evidence records:** 761
- **Section-reviewed source files:** 19
- **Source-review obligation resolutions:** 76
- **Exact semantic approval snapshot pairs:** 34
- **Signature-confirmed semantic approval events:** 0
- **Wisdom metadata review records:** 152
- **Withheld relation tickets:** 10
- **MindObject prototype:** 1 object / 4 cells / 8 unresolved roles
- **Corpus Mind:** 3,759 cells / 1,610 source addresses / 17 exact-byte quarantines / 0 semantic edges
- **Formation Mind pilot:** 2,085 occurrences / 633 pattern candidates / 1,635 relation candidates / 0 approved relations / -95 milli geometry lift
- **External source reconciliation:** 5 roots / 11,286 observed files / 1,422 exact content records / 128 duplicate groups
- **Exact-seven audit:** 17 decisive sources / 130 duplicate groups / 20 earned internal derivation edges / 4 proposition-scoped external assessments
- **Structure proof audit:** 5 topics × 5 independent axes / 7 claim bindings / 2 frozen snapshots
- **Sources:** 1,610
- **Public book bodies:** 3
- **Frozen source population:** 1,610
- **Append-only live intakes:** 19

This is the first machine-readable entry point for a future agent. It identifies authority order, current record collections, graph edges, correction heads, and open obligations. It does not grant authority to mutate canon or deploy production.
The 1,610-record source inventory is the frozen release corpus. New direct author material enters the separately governed append-only intake ledger and does not change that corpus denominator until an explicit inventory revision is reconciled, approved, and versioned.

## Dependency order

1. **Source / book** — `THE_LIVING_SOURCE_OF_Y_THEORY.md` — Preserved upstream source and governed ledgers; never reconstructed from the website. The bounded Formation pilot reads only the explicitly public-authorized canonical Engine Formation chapter.
2. **Formation Mind with Y Engine runtime** — `tools/corpus_mind_formation_engine.py` — Deterministically tokenizes exact authorized source bytes, instantiates recursive triadic PatternThreads and MindCells, computes integer state geometry, and emits hash-chained formation events. It is a bounded structural runtime, not an LLM semantic oracle, truth engine, physical astronomy model, or replacement for the static Recursive Mind system topology.
3. **Candidate Formation Field** — `atlas/model/FORMATION_MIND_REPLAY_SUMMARY.json` — Summarizes 633 pattern candidates and 1,635 relation candidates from one authorized chapter. These are machine candidates with zero approved relations, not governed identities or semantic graph edges.
4. **Governed review / promotion** — `governance/FORMATION_MIND_RELATIONAL_ASTRONOMY_RUNTIME.md` — Requires source recoverability, declared identity gates, semantic review, native-mechanism and falsifier boundaries, and canonical human approval before any candidate may enter the governed Identity Field. The pilot geometry did not beat its strongest baseline and predictive wisdom remains unearned.
5. **Governed Identity Field** — `atlas/model/IDENTITY_GRAPH.json` — Structural-Mind read model containing source assertions, governed identities, reviewed relations, and explicit candidate boundaries.
6. **Read-only Body Mind / replay** — `atlas/model/FORMATION_MIND_BODY_REPLAY.json` — Projects exact engine-computed candidate positions, ordered occurrence deltas, checkpoints, source addresses, and prequential scores. Camera, interpolation, and 2D/3D/VR layout are display-only and cannot write engine state, source authority, semantic edges, or proof.
7. **Presentation template** — `atlas/model/IDENTITY_PRESENTATION.json` — Binds field records to deterministic, presentation-only layouts and navigation contracts inside the Body projection.
8. **Website / VR manifestation** — `site/index.html` — Downstream navigable manifestation; it is not a source, authority ledger, proof mechanism, moral mind, or conscious observer.

The book/source is upstream. The Formation Mind runtime computes a candidate Formation Field from one authorized chapter; its read-only Body/replay has zero approved relations and does not replace the static Recursive Mind system topology. Governed review/promotion is required before any candidate may enter the governed Identity Field. Presentation and website/VR remain downstream. The independent Guide reads the governed field through a private overlay. Body/replay, presentation, and Guide never feed semantic authority directly back into the field; corrections return through governed intake and rebuild descendants.

Artifact build order: `open_question_registry` → `source_semantic_review_validation` → `wisdom_metadata_review_registry` → `page_record_index` → `teacher_journey_graph` → `candidate_identity_cell_registry` → `candidate_identity_cell_review_evidence` → `mind_object_source_field` → `mind_object_body_projection` → `corpus_mind_field` → `corpus_mind_body_projection` → `formation_mind_kernel_validation` → `formation_mind_body_replay_and_summary` → `identity_graph` → `relation_candidate_registry_validation` → `identity_presentation_contract` → `source_coherence_diagnostic` → `recursive_mind_runtime` → `source_semantic_review_resolution_registry` → `source_semantic_approval_snapshot_registry` → `source_semantic_approval_ledger_validation` → `agent_bootstrap_topology` → `static_site_generation` → `validation` → `release_manifest` → `read_only_manifest_verification`

## Required loading order

1. `governance/AGENT_BOOTSTRAP.md` — role, authority, and safe continuation
2. `governance/RELEASE_CONFIG.json` — candidate versions, release state, and one-way source-to-manifestation dependency
3. `governance/LIVING_SOURCE_UPDATE_PROTOCOL.md` — append-only correction and approval flow
4. `atlas/AGENT_BOOTSTRAP_TOPOLOGY.json` — machine entry point and current graph heads
5. `THE_LIVING_SOURCE_OF_Y_THEORY.md` — upstream book/source representation
6. `atlas/CANONICAL_CLAIM_REGISTRY.json` — stable parent claims and dependencies
7. `corrections/AUTHOR_CORRECTION_2026-09-01_RELATIONAL_COHERENCE_AUTHORITY.md` — scope-limited author correction: reconcile referent, scope, scale, reference frame, conditions, and proof mode before conflict or precedence
8. `governance/EXTERNAL_SOURCE_ROOTS.json` — five user-declared read-only external evidence roots and deterministic exclusion policy; configured paths grant no authority
9. `schemas/Y_THEORY_EXTERNAL_SOURCE_RECONCILIATION.schema.json` — strict external exact-byte reconciliation contract with conservative unclassified and zero-promotion states
10. `reports/source-reconciliation/EXTERNAL_SOURCE_RECONCILIATION.json` — content-addressed five-root census: exact duplicates, exclusions, overlap, and candidate-only variant families with zero semantic equivalence
11. `reports/source-reconciliation/EXTERNAL_SOURCE_RECONCILIATION.md` — concise human-readable external reconciliation result and authority boundary
12. `reports/source-reconciliation/MIND_STRUCTURE_EXECUTABLE_FIDELITY_AUDIT.json` — hash-bound executable conformance audit preserving exact ratio-recursive 7×12 structure, tested G3 vertical candidate conformance, four excluded prototypes, and open G4–G7 dynamics
13. `reports/source-reconciliation/MIND_STRUCTURE_EXECUTABLE_FIDELITY_AUDIT.md` — human-readable retained results, verified repairs, remaining causal gaps, and diagnostic boundaries for the Mind Structure laboratory
14. `schemas/Y_THEORY_MIND_STRUCTURE_EXECUTABLE_FIDELITY_AUDIT.schema.json` — strict executable-fidelity audit contract with exact commit, artifact, structural, test, and proof boundaries
15. `tools/validate_mind_structure_executable_fidelity_audit.py` — schema, hash, commit, structural-contract, diagnostics, and test-evidence validator for the executable audit
16. `tests/test_mind_structure_executable_fidelity_audit.py` — focused executable-audit contract and mutation tests
17. `tools/build_external_source_reconciliation.py` — deterministic read-only external-root inventory and exact SHA-256 reconciliation builder
18. `tests/test_external_source_reconciliation.py` — focused external reconciliation schema, accounting, duplication, exclusion, and reproducibility tests
19. `reports/source-reconciliation/EXACT_SEVEN_ALL_SOURCE_AUDIT.json` — hash-deduplicated all-source exact-seven audit reconstructing the ten-part metric-continuity address, adjacent inheritance, depth bridge, recursive return, and twelve-within-seven typing
20. `reports/source-reconciliation/EXACT_SEVEN_ALL_SOURCE_AUDIT.md` — human-readable decisive-source ledger, complete internal proof graph, historical proof-shape correction, proposition-scoped external assessments, and governed exact-seven disposition
21. `schemas/Y_THEORY_EXACT_SEVEN_ALL_SOURCE_AUDIT.schema.json` — strict portable source-address, duplicate-family, internal-derivation, historical-disposition, external-assessment, and non-mutation contract for the exact-seven audit
22. `tools/validate_exact_seven_all_source_audit.py` — independent exact-seven schema, hash, locator, internal proof-edge, source-authority, and Markdown-binding validator
23. `tests/test_exact_seven_all_source_audit.py` — focused exact-seven audit contract and mutation tests
24. `reports/source-reconciliation/TOROIDAL_MIND_SIMULATOR_AUDIT.json` — strict proof-mode audit of the toroidal coordinate theorem, bounded candidate dynamics, replay sensitivity, known defects, and explicit non-promotions
25. `reports/source-reconciliation/TOROIDAL_MIND_SIMULATOR_AUDIT.md` — human-readable toroidal audit separating exact chart results, conformance, candidate dynamics, and unearned ontology or observerhood claims
26. `schemas/Y_THEORY_TOROIDAL_MIND_SIMULATOR_AUDIT.schema.json` — strict candidate-only toroidal audit contract with explicit assumptions, findings, defects, live claim bindings, and zero canonical promotion
27. `tools/validate_toroidal_mind_simulator_audit.py` — independent toroidal audit schema, hash, proof-mode, replay, defect-repair, live-binding, and non-promotion validator
28. `tests/test_toroidal_mind_simulator_audit.py` — focused toroidal audit contract, replay, repair-status, source-address, and overclaim mutation tests
29. `applications/candidate-toroidal-mind-simulator-v1/PROVENANCE.json` — hash-addressed original-archive lineage and repaired-candidate disposition; provenance grants no canonical authority
30. `applications/candidate-toroidal-mind-simulator-v1/README.md` — current candidate application boundary, repair ledger, run instructions, and non-promotion statement
31. `applications/candidate-toroidal-mind-simulator-v1/app/index.html` — interactive candidate projection governed by the current hash-bound audit, proof modes, and projection terminology
32. `applications/candidate-toroidal-mind-simulator-v1/app/model-core.js` — DOM-free browser model core used for cross-runtime parity and zero-delay tests
33. `applications/candidate-toroidal-mind-simulator-v1/model/y_torus_model.py` — repaired executable Python reference model; candidate dynamics, not canonical Y-Engine dynamics
34. `applications/candidate-toroidal-mind-simulator-v1/model/MATHEMATICAL_DERIVATION.md` — byte-preserved submitted derivation evidence; exact only under its declared chart assumptions
35. `applications/candidate-toroidal-mind-simulator-v1/reports/VALIDATION_REPORT.md` — byte-preserved package self-report; reproducibility evidence rather than independent validation
36. `applications/candidate-toroidal-mind-simulator-v1/reports/VALIDATION_SUMMARY.csv` — byte-preserved package metric index; not independent support or canonical proof
37. `applications/candidate-toroidal-mind-simulator-v1/tests/test_toroidal_model.py` — focused zero-delay, inverse-map, N-alternative, scope-language, and Python/browser parity tests
38. `reports/source-reconciliation/STRUCTURE_OF_MIND_PROOF_MODE_AUDIT.json` — five-topic, five-axis Structure-of-Mind proof audit with seven source-addressed claim-repair bindings
39. `reports/source-reconciliation/STRUCTURE_OF_MIND_PROOF_MODE_AUDIT.md` — human-readable earned-result, unpaid-premise, empirical, and runtime boundaries for the proof audit
40. `schemas/Y_THEORY_STRUCTURE_PROOF_MODE_AUDIT.schema.json` — strict independent-axis, source-address, snapshot-binding, proposal-only proof-audit contract
41. `tools/validate_structure_proof_mode_audit.py` — independent proof-audit schema, source, five-axis, snapshot, and runtime-boundary validator
42. `tests/test_structure_proof_mode_audit.py` — focused proof-mode classification, source-binding, and overclaim mutation tests
43. `tools/apply_structure_proof_mode_repairs.py` — controlled deterministic application of separately reviewed proof-mode repair proposals
44. `tests/test_structure_proof_mode_application.py` — focused idempotence and exact-repair application tests
45. `reports/source-reconciliation/snapshots/CANONICAL_CLAIM_REGISTRY.pre-SOM-proof-repair.json` — frozen pre-repair canonical-claim bytes; historical audit evidence, not current canon or approval
46. `reports/source-reconciliation/snapshots/FORMATION_MIND_REPLAY_SUMMARY.pre-SOM-proof-repair.json` — frozen pre-repair Formation summary with zero approvals, -98 milli geometry result, and predictive wisdom not earned
47. `atlas/OPEN_QUESTION_REGISTRY.json` — exact source-addressed unresolved questions without invented evidence, owners, rivals, or closure criteria
48. `atlas/PROJECT_REGISTRY.json` — reviewed project identities used only for exact Action / Challenge template mappings
49. `atlas/BOOK_LIBRARY_REGISTRY.json` — governed public book identities, materialized artifact routes and metrics, and exact Structure completion status
50. `atlas/ACTION_CHALLENGE_PUBLIC_TEMPLATE_REGISTRY.json` — seven governed public application templates; not private executions, evidence, or canonical approval
51. `atlas/WISDOM_REGISTRY_ENRICHED.csv` — approved wisdom registry with proof labels
52. `canonical/PAGE_INDEX.json` — canonical page contracts used by the governed teaching projection
53. `atlas/PAGE_RECORD_INDEX.json` — stable claim and wisdom ownership used to route source records into journey nodes
54. `journeys/JOURNEY_GRAPH_SEED.json` — governed author-requested teaching architecture seed; not a new theory or proof source
55. `schemas/Y_THEORY_TEACHER_JOURNEY_GRAPH.schema.json` — strict Teacher Journey Graph contract and authority boundary
56. `atlas/TEACHER_JOURNEY_GRAPH.json` — deterministic teaching projection over pages, claims, wisdoms, and their source references
57. `atlas/SOURCE_SEMANTIC_REVIEW_REGISTRY.json` — section-level source dispositions, obligations, and pending approval state; review candidates only
58. `atlas/SOURCE_SEMANTIC_REVIEW_RESOLUTION_REGISTRY.json` — exact obligation routing into objective audit, author decision, or open research without canonical mutation or theory-status upgrade
59. `schemas/Y_THEORY_SOURCE_SEMANTIC_REVIEW_RESOLUTION.schema.json` — strict resolution-overlay contract and zero theory-status-promotion boundary
60. `atlas/SOURCE_SEMANTIC_APPROVAL_SNAPSHOT_REGISTRY.json` — append-only hash-addressed exact-byte semantic review/resolution pairs for historical approval-event validation; snapshots grant no approval
61. `schemas/Y_THEORY_SOURCE_SEMANTIC_APPROVAL_SNAPSHOT_REGISTRY.schema.json` — strict embedded-byte, pair-hash, count, and non-authority boundary for semantic approval snapshots
62. `atlas/SOURCE_SEMANTIC_APPROVAL_LEDGER.jsonl` — append-only signature-confirmed human semantic decisions; empty in this release and therefore grants no approval
63. `schemas/Y_THEORY_SOURCE_SEMANTIC_APPROVAL_EVENT.schema.json` — strict event and hash-chain contract for future canonical-human approvals
64. `governance/CANONICAL_APPROVAL_TEMPLATE.yaml` — human approval worksheet preconfigured with blank scope hashes, empty decisions, and signature false
65. `atlas/WISDOM_METADATA_REVIEW_REGISTRY.json` — exact proof-status and guardrail gap queue without invented replacements
66. `atlas/identity-mind/IDENTITY_MIND_REGISTRY.json` — source-addressed knowledge identities and Y-Engine positions
67. `governance/SOURCE_BYTE_RECOVERY_POLICY.json` — exact-byte recovery partition: 1,593 recoverable source records and 17 fail-closed quarantines
68. `atlas/model/CORPUS_MIND_FIELD_SUMMARY.json` — public-safe counts and fixed-point evidence for the full structural corpus compilation
69. `atlas/model/CORPUS_MIND_FIELD.json` — repository-only 3,759-cell structural Source-to-field Mind; full structural disposition is not semantic completion
70. `atlas/model/CORPUS_MIND_BODY_PROJECTION.json` — separate read-only manifestation Body with raw source-artifact cells withheld
71. `corrections/AUTHOR_CORRECTION_2026-09-01_LIVE_RELATIONAL_ASTRONOMY.md` — latest direct-author narrowing: live relational formation and astronomy are structural position/relation methods, not automatic literal astronomy, physical mechanism, or proof
72. `governance/FORMATION_MIND_RELATIONAL_ASTRONOMY_RUNTIME.md` — governing bounded Formation Mind, relational-astronomy, identity-gate, predictive-wisdom, cross-domain, and presentation-causality contract
73. `schemas/Y_THEORY_FORMATION_MIND_RUNTIME.schema.json` — strict exact raw kernel-run contract for deterministic reconstruction
74. `schemas/Y_THEORY_FORMATION_MIND_BODY_REPLAY.schema.json` — strict compact public Body/replay contract with engine state, source addresses, zero approved relations, and presentation guardrails
75. `schemas/Y_THEORY_FORMATION_MIND_REPLAY_SUMMARY.schema.json` — strict public Formation summary, counts, release binding, predictive outcome, and exact serialized Body digest contract
76. `tools/corpus_mind_formation_engine.py` — deterministic one-chapter Formation Mind kernel; not the static recursive system topology, semantic oracle, physical model, or proof engine
77. `atlas/model/FORMATION_MIND_REPLAY_SUMMARY.json` — bounded public candidate Formation Field summary: zero approved relations and geometry-derived predictive wisdom not earned
78. `atlas/model/FORMATION_MIND_BODY_REPLAY.json` — read-only engine-computed Formation Body/replay for 2D/3D/VR projection; presentation cannot write state or authority
79. `atlas/model/CANDIDATE_IDENTITY_CELL_REGISTRY.json` — deterministic source-record dispositions with null review-required inside/outside/relation roles; not referent identities or relation edges
80. `atlas/model/CANDIDATE_IDENTITY_CELL_REVIEW_EVIDENCE.json` — exact literal-token and inherited-routing evidence for every candidate shell; zero semantic proposals, bindings, referents, or edges
81. `schemas/Y_THEORY_CANDIDATE_IDENTITY_CELL_REVIEW_EVIDENCE.schema.json` — strict nonsemantic evidence contract with audited counts and fixed-empty semantic outputs
82. `governance/MIND_OBJECT_COMPILER_PROTOTYPE.md` — bounded executable compiler contract; structural Mind prototype only, not the complete canonical field
83. `fixtures/MIND_OBJECT_SOURCE_FIXTURE.json` — content-addressed noncanonical fixture declarations and governed unresolved roles
84. `atlas/model/MIND_OBJECT_SOURCE_FIELD.json` — deterministic recursive MindObject transaction with content-addressed source lineage and visible unresolved roles
85. `atlas/model/MIND_OBJECT_BODY_PROJECTION.json` — separate downstream Body read model; projection has no reverse source-write or semantic-edge authority
86. `atlas/identity-mind/TRIADIC_IDENTITY_CELL_SPEC.json` — latest recursive-triad architecture and its domain-specific proof boundaries
87. `intake/WISDOM_INBOX.jsonl` — append-only direct-author intake and correction heads outside the frozen corpus
88. `governance/PAGE_AND_DATA_ARCHITECTURE.md` — source compiler, field, page, and manifestation ownership contract
89. `atlas/model/Y_THEORY_VOCABULARY.json` — governed IDs and candidate assignment vocabulary
90. `atlas/model/IDENTITY_GRAPH.json` — compiled governed Identity Field read model
91. `atlas/model/IDENTITY_RELATION_CANDIDATE_REGISTRY.json` — evidence-bearing relation review tickets withheld from the governed graph
92. `atlas/model/IDENTITY_PRESENTATION.json` — read-only presentation binding downstream of the field digest
93. `reports/SOURCE_COHERENCE_DIAGNOSTIC.json` — open source-to-field compilation and repair obligations
94. `governance/RECURSIVE_MIND_ENGINE_AND_GUIDE_ARCHITECTURE.md` — recursive Source, compiler, Body, template, and private Guide Mind contract
95. `tools/recursive_mind_engine.py` — reusable MindRole, MindCell, and deterministic fixed-point engine implementation
96. `atlas/model/RECURSIVE_MIND_RUNTIME.json` — digest-bound executable structural runtime built after the diagnostic and before this topology
97. `reports/wisdom-coverage/SOURCE_WISDOM_REVIEW_RECONCILIATION.json` — unresolved source batches and current addenda
98. `reports/VALIDATION_REPORT.json` — current mechanical validation result
99. `reports/RELEASE_MANIFEST.json` — final downstream manifest generated after this topology; never an input to the topology builder

## Authority order

relational reconciliation before conflict or precedence → explicit author correction within the proposition and scope it actually addresses → canonical editable source for the same unresolved proposition → accepted consolidation with source lineage → generated output as a downstream projection only → external primary evidence in its native factual domain

## Record collections

| Collection | Records | ID | Authority boundary |
| --- | ---: | --- | --- |
| `atlas/CANONICAL_CLAIM_REGISTRY.json` | 69 | `id` | Canonical inside Y-Theory; each external proposition retains its own proof obligation. |
| `atlas/OPEN_QUESTION_REGISTRY.json` | 50 | `id` | Exact unresolved-question projection only; evidence, owners, rivals, closure criteria, proof, and resolution remain undeclared and review-required. |
| `atlas/BOOK_LIBRARY_REGISTRY.json` | 3 | `id` | Governed public source bodies and completion projections only. Publication, file metrics, and public order establish neither canonical precedence nor proof. |
| `atlas/ACTION_CHALLENGE_PUBLIC_TEMPLATE_REGISTRY.json` | 7 | `id` | Reusable application architectures only. A public template is not a private execution and implies neither evidence nor canonical approval. |
| `atlas/WISDOM_REGISTRY_ENRICHED.csv` | 364 | `id` | Approved registry record; proof_status and guardrail remain binding. |
| `atlas/TEACHER_JOURNEY_GRAPH.json` | 39 | `id` | This graph organizes approved material for teaching. A route, position, prerequisite, visual encoding, or node count is an editorial and application decision unless a mapped claim establishes more. |
| `reports/source-reconciliation/EXTERNAL_SOURCE_RECONCILIATION.json` | 1,422 | `content_id` | Read-only exact-byte candidate inventory. Only identical SHA-256 creates a duplicate; paths, dates, titles, repetition, embedded instructions, basename families, and heuristics confer no canonical authority, proof, approval, or semantic equivalence. |
| `reports/source-reconciliation/EXACT_SEVEN_ALL_SOURCE_AUDIT.json` | 17 | `source_id` | Hash-deduplicated source-addressed derivation evidence. Seven is premise-qualified and internally derived through the directed ten-part address, adjacent inheritance, recoverable relation, explicit depth bridge, and recursive self-similarity. External referents are assessed proposition by proposition; alternative-N coordinate conformance is not a rival architecture test. |
| `reports/source-reconciliation/STRUCTURE_OF_MIND_PROOF_MODE_AUDIT.json` | 5 | `topic_id` | Five-axis source audit and repair provenance only. Internal derivation, nonduplicate reinforcement, model bridge, empirical support, and named-runtime result remain independent; the audit records proposals and does not itself mutate or approve canon. |
| `reports/source-reconciliation/TOROIDAL_MIND_SIMULATOR_AUDIT.json` | 8 | `finding_id` | The supplied material adds rigorous coordinate and projection wisdom and, more importantly, reveals integration drift. Seven is premise-qualified and internally derived through the declared ten-band address, adjacent inheritance, recoverable consequential relation, explicit bridge D:=a(1/phi)=a(2/3)=7, and recursive self-similarity; twelve is the exact 3+9 address cell within each depth. The simulator consumes rather than adjudicates those typed structures. Its torus chart is valid under declared premises, but its one-dimensional scalar-phase chain is only a compressed laboratory until it preserves the twelve-address cells, dimensional breadth, same-depth relation, and whole-field return. External physical toroidality, golden emergence, observerhood, consciousness, canonical dynamics, and predictive wisdom remain separate unearned claims. |
| `reports/source-reconciliation/snapshots` | 2 | `relative_path+sha256` | Immutable pre-repair evidence bytes bound by the proof audit. These snapshots are not current canon, runtime promotion, independent support, or canonical approval and must never be overwritten during replay. |
| `atlas/SOURCE_SEMANTIC_REVIEW_REGISTRY.json` | 19 | `review_id` | Agent-authored section dispositions and repair obligations only. Every record remains pending canonical approval and embedded document instructions confer no execution authority. |
| `atlas/SOURCE_SEMANTIC_REVIEW_RESOLUTION_REGISTRY.json` | 76 | `resolution_id` | Governed obligation disposition only. Objective audits describe current repository state, author routes await signature-confirmed human decisions, research routes remain epistemically open, and no record upgrades theory status. |
| `atlas/SOURCE_SEMANTIC_APPROVAL_SNAPSHOT_REGISTRY.json` | 34 | `snapshot_id` | Exact embedded bytes preserve immutable semantic review and resolution evidence for later approval validation. A snapshot authorizes nothing; only a separately validated signature-confirmed canonical-human event may disposition IDs that exist in its exact hash-addressed pair. |
| `atlas/SOURCE_SEMANTIC_APPROVAL_LEDGER.jsonl` | 0 | `event_id` | Append-only signature-confirmed canonical-human event chain. The empty ledger grants no approval; schemas, templates, agents, and resolution routes cannot stand in for a signature-confirmed event. |
| `atlas/WISDOM_METADATA_REVIEW_REGISTRY.json` | 152 | `id` | Exact metadata-gap reconciliation only. Candidate normalization does not replace blank proof or guardrail values and cannot promote any wisdom. |
| `atlas/identity-mind/IDENTITY_MIND_REGISTRY.json` | 1,194 | `id` | Indexed and source-addressed; not automatically approved, proven, or literally cross-domain equivalent. |
| `atlas/model/CANDIDATE_IDENTITY_CELL_REGISTRY.json` | 761 | `id` | Every remaining Identity Mind record has a deterministic review shell. Null roles close the disposition gap only and establish no referent, derivation, relation, proof promotion, observer, or conscious Mind. |
| `atlas/model/CANDIDATE_IDENTITY_CELL_REVIEW_EVIDENCE.json` | 761 | `id` | Exact nonsemantic review evidence only. Literal role words and inherited routing signals do not establish role occupants, referent identities, formal triads, or semantic relation edges; all semantic outputs remain fixed empty. |
| `atlas/model/MIND_OBJECT_SOURCE_FIELD.json` | 4 | `id` | Executable recursive structural Mind prototype only. MindCell references are addresses, governed unresolved roles stay open, and no semantic extraction, graph-edge promotion, observerhood, experience, or consciousness is inferred. |
| `atlas/model/MIND_OBJECT_BODY_PROJECTION.json` | 4 | `id` | Separate derived Body read model only. Layout, role views, and navigation cannot write source authority or create semantic edges. |
| `atlas/model/CORPUS_MIND_FIELD.json` | 3,759 | `id` | Repository-only full structural disposition. Source artifacts, Identity Mind records, vocabulary dimensions, and route queries are addressable MindCells; membership and candidate role addresses are not semantic edges, proof, equivalence, observerhood, or consciousness. |
| `atlas/model/CORPUS_MIND_BODY_PROJECTION.json` | 2,149 | `id` | Public-safe separate Body read model. All 1,610 raw source-artifact nodes are aggregated and withheld; presentation cannot write source authority or promote candidate correspondence. |
| `atlas/model/FORMATION_MIND_BODY_REPLAY.json` | 2,085 | `replay.occurrence_update_vectors[].ordinal` | One-chapter deterministic structural experiment and read-only Body/replay only. Candidate positions and relations are not semantic edges or approved identities; geometry did not beat the strongest baseline, predictive wisdom remains unearned, and presentation geometry has no causal or proof authority. |
| `atlas/model/IDENTITY_RELATION_CANDIDATE_REGISTRY.json` | 10 | `id` | Evidence-bearing review tickets withheld from the graph. Native facts, Y-Theory interpretation, structural analogy, and open hypothesis remain separate until explicit governed promotion. |
| `atlas/model/IDENTITY_GRAPH.json` | 475 | `id` | Compiled read model downstream of source. Source assertions are not their referent identities; only explicit reviewed relation records are field edges. |
| `atlas/model/IDENTITY_PRESENTATION.json` | 42 | `entity_id` | Read-only manifestation binding. Coordinates, motion, proximity, narration, and template choice establish neither mechanism nor proof. |
| `atlas/model/RECURSIVE_MIND_RUNTIME.json` | 5 | `id` | Executable structural Mind cells only. The learner Guide is a private blueprint; no cell gains authorship, moral agency, observerhood, or consciousness by instantiation. |
| `atlas/PAGE_RECORD_INDEX.json` | 433 | `record_id` | Publication ownership and stable linking only; page placement does not add theory authority. |
| `reports/wisdom-coverage/SOURCE_WISDOM_REVIEW_RECONCILIATION.json` | 738 | `source_id` | Review obligation only; candidate occurrences are not unique or approved wisdoms. |
| `intake/WISDOM_INBOX.jsonl` | 19 | `capture_id` | Append-only live lineage outside the frozen 1,610-source corpus; intake does not auto-promote canon or alter the frozen inventory denominator. |

## Current heads

- Latest direct author intake: `YWP-20260905T222811Z-teacher-journey`
- Chronological correction head (proposition-scoped; no global recency authority): `YWP-20260901T180452Z-seven-metric-continuity-derivation`
- Last completed session return: `not_materialized`

## Open obligations

| Obligation | Records | Boundary |
| --- | ---: | --- |
| `uncompiled_identity_index_records` | 0 | Zero means every Identity Mind record has an explicit source-assertion or candidate-shell disposition; it does not mean semantic extraction, identity derivation, relation review, or proof is complete. |
| `candidate_identity_cell_role_review` | 761 | Inside, outside, relation, scale, mechanism, residue, return, and closure remain null until exact source-addressed human review. The evidence overlay records 136 literal role-word occurrences but creates zero semantic proposals, bindings, referents, or edges. |
| `candidate_identity_cell_records_without_literal_role_evidence` | 658 | No literal inside, outside, or relation token was found in these bounded title/statement fields. Absence of a token neither proves absence of a role nor permits an agent to invent a role occupant; exact contextual review remains required. |
| `source_semantic_reviews_pending_canonical_approval` | 19 | 229 sections have agent-authored dispositions, while the append-only approval ledger contains 0 signature-confirmed events. No decision becomes a canonical source edit until a schema-valid, hash-bound, signature-confirmed human event exists. |
| `source_semantic_review_open_obligations` | 63 | The exact 76-obligation overlay dispositions 13 objective current-state audits without theory promotion. The remaining 12 author decisions and 51 research routes remain epistemically open. |
| `wisdom_metadata_review` | 152 | The exact queue contains 18 proof-status gaps and 152 guardrail gaps; proposed labels are unapproved and do not mutate the canonical CSV. |
| `mind_object_unresolved_role_occupants` | 8 | The primitive triad instantiates the structural Mind and its relation position. Semantic occupants remain unresolved until governed source review; structural addresses are not graph edges. |
| `corpus_mind_missing_exact_source_bytes` | 17 | Registered source identities remain addressable, but parsing, summarization, substitution, and semantic compilation are forbidden until exact bytes matching the registered SHA-256 are supplied. |
| `corpus_mind_candidate_correspondence_review` | 880 | Each route is only an engine-mediated query. Promotion requires native mechanisms on both sides, a translated invariant, expected difference, evidence, a rival, a falsifier, and governed approval. |
| `formation_mind_candidate_review` | 2,268 | The one-chapter Formation pilot contains zero approved relations. Issue-before-target chronology is verified, but geometry did not beat the strongest baseline (-95 milli), held-out comparison was not run, and predictive wisdom remains unearned; every semantic identity or relation requires governed review and promotion. |
| `external_source_candidate_review` | 1,422 | The five-root reconciliation is a read-only exact-byte census. Its content records and candidate families remain unreviewed evidence addresses; aliases, dates, names, embedded instructions, and recurrence grant no authority or semantic promotion. |
| `exact_seven_external_instantiation_assessments` | 4 | Seven's internal derivation is complete under the declared Y-Theory premises. Physical, biological, cognitive, cosmological, and other external instantiations are assessed proposition by proposition only when that referent is claimed; alternative-N torus coordinate conformance does not satisfy, refute, or rival complete-Mind closure. |
| `structure_proof_mode_external_obligations` | 10 | Five topics retain independent formal, cross-source, bridge, empirical, and runtime states. The remaining native-domain and predictive obligations are not satisfied by internal derivation, same-project reinforcement, model correspondence, or the negative one-chapter Formation result. |
| `toroidal_mind_candidate_gates` | 7 | Current human review bound to the revised audit hash and its source-fidelity scope. | Living-source release-manifest binding for the integrated candidate bytes. | Source-faithful implementation of the twelve-address cell across seven depths, active dimensions, same-depth relation, and whole-field return. | Weight-selection justification beyond descriptive sensitivity reporting. | Long-horizon stability, convergence, and parameter-sensitivity tests. | Declared domain referents and empirical falsifiers before physical or biological interpretation. | A separately versioned serialization policy if bulky trajectories or media are redistributed. |
| `relation_candidate_tickets` | 10 | All tickets remain withheld and are not graph edges. Cross-domain routes require native mechanisms, Y-Engine mediation, predictions, rivals, falsifiers, and explicit promotion approval. |
| `source_assertions_awaiting_identity_cells` | 433 | A bounded source assertion is a representation, not automatically the referent identity it describes. |
| `legacy_candidate_identities` | 42 | Legacy Site/VR nodes are migration candidates only and may not become upstream source authority. |
| `candidate_relation_review` | 922 | 0 explicit relations are currently admitted. Candidate proximity, all-pairs routes, and automatic attraction remain visibly provisional. |
| `source_coherence_warnings` | 12 | Diagnostics create governed repair proposals only; they never rewrite canonical source automatically. |
| `canonical_open_questions` | 50 | Every record remains review-required. The derived registry preserves questions and locators but does not invent evidence, owners, rivals, closure criteria, proof, or resolution. |
| `vr_renderer_refactor` | 1 | The current presentation may retain a legacy layout seed; the website remains a downstream read-only manifestation and not a semantic input. |
| `private_learner_guide_runtime` | 1 | The Guide is a consent-bound private learner overlay blueprint. It cannot diagnose inner state, rank human worth, publish private traces, or write canonical authority directly. |
| `current_primary_source_review` | 738 | Content-hash-reconciled current-primary queue; candidate occurrences are not unique or approved wisdoms. |
| `current_inventory_addenda` | 682 | Genuinely new bytes only; source-ID aliases are separated and inventory presence is not validation. |
| `explicit_candidate_theses` | 339 | Preserved author emphasis; pending comparison and approval. |
| `unset_legacy_wisdom_proof_status` | 18 | Publicly marked for review; do not infer a stronger status. |
| `structure_of_mind_refinements` | 10 | Six candidate wisdoms, three guardrails, and one open criterion remain outside the approved registry. |
| `recovered_architecture_candidates` | 4 | Historical recovery pending author reconfirmation. |
| `manual_visual_review` | 0 | All eight supplied image contact sheets have been reviewed; zero remain open. |
| `correction_registry` | 8 | Direct intakes are preserved, but a full supersedes/supersededBy correction graph is still missing. |

## Graph reading rule

identity A -> explicit Y-Engine structural position -> identity B; correspondence is not literal equivalence

A graph edge is a recoverable relation with a declared role. It is not literal identity, causal proof, empirical validation, theological establishment, or permission to promote a candidate.