AdaWorldAPI · AdaWorldAPI · May 12, 2026 · May 12, 2026 · May 12, 2026 · May 12, 2026
diff --git a/.claude/board/INTEGRATION_PLANS.md b/.claude/board/INTEGRATION_PLANS.md
@@ -345,3 +345,45 @@ These existing plans absorb cleanly into the new architecture and remain in scop
 ### Sprint-2 governance
 
 This sprint was orchestrated as 12 worker agents + 1 meta agent on branch `claude/unified-ogit-architecture-synthesis`. CCA2A pattern: per-agent append-only logs in `.claude/board/sprint-log-2/agents/agent-W*.md`; meta review in `.claude/board/sprint-log-2/meta-1-review.md`; sprint summary in `.claude/board/sprint-log-2/sprint-summary.md`.
+
+## 2026-05-12 — Sprint-3: Tier-1 Implementation Specs (PR #360 + #361 + substrate-recognition sweep)
+
+11 PR-X-1 implementation specs landed via PR #360 + 2 architectural corrections via PR #361 + 3 spec re-scopes via post-#360 substrate-sweep PR. After this sprint sequence, an engineer picks any PR-X-Y spec and starts coding without re-design.
+
+### New specs (`.claude/specs/`)
+
+| Spec | Pattern | Status | Effort (post-substrate-sweep) |
+|---|---|---|---|
+| `sprint-3-execution-plan.md` | (master) | Active | n/a |
+| `sprint-3-pr-graph.md` | (sequencing; compressed timeline) | Active | n/a |
+| `pr-a-1-spo-g-u32-slot.md` | A | Active (re-scoped) | ~150 LOC / 1 day |
+| `pr-b-1-context-bundle.md` | B | Active | ~200 LOC / 1 day |
+| `pr-c-1-generic-bridge.md` | C | Active (re-scoped) | ~80 LOC / ½ day |
+| `pr-d-1-fma-owl-hydrator.md` | D (PARTIALLY SHIPPED) | Active (re-scoped) | ~250 LOC / 1-2 days |
+| `pr-e-1-manifest-modules.md` | E | Active (post-#361 cycle fix) | ~330 LOC / 2 days |
+| `pr-f-1-ractor-supervisor.md` | F | Active (post-#361 inert-bundle skip) | ~400 LOC / 3 days |
+| `pr-j-1-int4-32d-atoms.md` | J | Active | ~120 LOC / 1 day |
+| `consumer-crate-template.md` | (Pattern C dry-run; re-targeted to woa-rs + medcare-rs precedents) | Active | n/a |
+| `ogit-g-smoke-test.md` | (validation; PR-A-1+B-1+C-1+E-1+F-1) | Active | ~200 LOC / 1 day |
+| `trivia-prs-bundle.md` | (3 quick wins: TD-CAM-DIST + TD-ADJ-THINK + TD-DEEPNSM-NSM) | Active | ~60 LOC total / <1 day |
+
+### New knowledge docs (`.claude/knowledge/`)
+
+- `pattern-recognition-cross-source.md` — 4-taxonomy matrix (A-O ↔ Pillars 0-4 ↔ `.grok/` ↔ shipped substrate)
+- `cca2a-sprint-prompt-template.md` — substrate-grep checklist + wrong-repo guardrail + pattern-letter discipline
+
+### Pre-existing plans absorbed by sprint-3
+
+- `lance-graph-ontology-v5.md` — Pillar 0 (already shipped via PR #355)
+- `palantir-parity-cascade-v2.md` — Pillars 0-4 architecture (parallel taxonomy; cross-referenced via new matrix)
+- `ogit-cascade-supabase-callcenter-v1.md` — Pillars 0-4 cascade execution (parallel taxonomy)
+- `unified-ogit-architecture-v1.md` (W1 master) — sprint-2 north star
+- `ogit-g-context-bundle-v1.md` (sprint-2 Tier-1 sub-plan)
+- `compile-time-consumer-binding-v1.md` (sprint-2 Tier-2 sub-plan)
+- `anatomy-realtime-v1.md` (sprint-2 proof-of-vision)
+
+### Cross-references
+
+- `.claude/board/LATEST_STATE.md` — sprint-3 entry (paired with this index update)
+- `.claude/board/TECH_DEBT.md` — TD-OGIT-G-SLOT-1 through TD-DEEPNSM-NSM-COLLAPSE-11 (the 11 entries spec'd by sprint-3)
+- `.claude/board/sprint-log-3/` — coordination directory (12 agent logs + meta-1 + sprint-summary)
diff --git a/.claude/board/LATEST_STATE.md b/.claude/board/LATEST_STATE.md
@@ -317,3 +317,41 @@ The master plan-doc `.claude/plans/unified-ogit-architecture-v1.md` provides the
 - **Risk:** the "~80% already shipped" claim is W1/W2's recognition assertion, not independently re-verified by W9. This section reports it as the synthesis output; the canonical evidence lives in `tier-0-pattern-recognition.md` (W2) and the entropy ledger reframe rows (W6).
 - **Governance:** append-only preserved. No deletions. No edits to the prior `## 2026-05-07 — Append: lance-graph-ontology shipped` section. Section heading matches the spec exactly.
 - **What this section does NOT do:** it does not edit the top-of-file "Last updated" line (would violate append-only); it does not edit the "Recently Shipped PRs" table (Sprint-2 shipped no PRs); it does not edit "Active Branches" (Sprint-2 is documentation tier on a branch that has not yet merged).
+
+## 2026-05-12 — Sprint-3: Tier-1 Implementation Specs (PR #360 + #361 + post-#360 substrate sweep)
+
+**PR #360** (sprint-3 main): 11 PR-X-1 specs covering 7 design-phase patterns A/B/C/D/E/F/J + 3 trivia closures + supporting docs. ~140 KB across `.claude/specs/`. Engineer can now execute Tier-1 in ~6 working days parallelized (per W10 sequencing).
+
+**PR #361** (post-#360 corrections): PR-F-1 supervisor must skip inert bundles (DOLCE/FMA have consumer_pointer=None by design); PR-E-1 build script must emit data-only (no consumer crate refs) to avoid Cargo dependency cycle. Both fixed via append-only correction sections; inventory-crate self-registration recommended for actor binding.
+
+**Post-#360 substrate-recognition sweep** (this PR): 3 of 11 specs reclassified PARTIALLY SHIPPED:
+- Pattern A: SchemaPtr.ontology_context_id + NamespaceRegistry::seed_defaults already ship; PR-A-1 reduces to ~150 LOC / 1 day
+- Pattern C: BridgeFromRegistry + 3 impls + woa-rs#2 + medcare-rs#110 consumer scaffolds already ship; PR-C-1 reduces to ~80 LOC / ½ day
+- Pattern D: parse_ttl_directory_with_provenance + attach_provenance already ship; PR-D-1 reduces to ~250 LOC / 1-2 days (OWL/RDF-XML adapter only)
+
+Compressed sprint-3 critical path: ~6 days → ~3-4 days parallelized. The genuinely-new ~5-pattern set is B (context bundle), E (manifest-modules), F (ractor port), G (inheritance protocol), J (INT4-32D atoms).
+
+### New knowledge docs (sprint-3 substrate-sweep)
+
+- `.claude/knowledge/pattern-recognition-cross-source.md` — A-O ↔ Pillars 0-4 ↔ `.grok/` ↔ shipped substrate matrix (4 parallel taxonomies cross-referenced)
+- `.claude/knowledge/cca2a-sprint-prompt-template.md` — substrate-grep checklist + wrong-repo guardrail + pattern-letter discipline (mandatory pre-spawn template for future sprints)
+
+### Anti-Pattern recurrence captured
+
+The "Designing What's Already Built" anti-pattern (introduced PR #358) recurred in sprint-3's own design (PR-A-1/PR-C-1/PR-D-1 over-scoped because they didn't sweep post-#355 substrate). The correction PR formalizes the substrate-grep checklist as mandatory before any new spec.
+
+### Recurring failure mode: wrong-repo error
+
+Sprint-2 W7 → ndarray; sprint-3 W9 → ada-consciousness. Both corrected via main-thread pygithub recovery. Wrong-repo guardrail snippet now mandatory in every worker prompt (per `.claude/knowledge/cca2a-sprint-prompt-template.md`).
+
+### Cross-references
+
+- `.claude/specs/sprint-3-execution-plan.md` (W1 master)
+- `.claude/specs/sprint-3-pr-graph.md` (W10 sequencing — to be updated for compressed timeline)
+- `.claude/specs/pr-{a,b,c,d,e,f,j}-1-*.md` (11 PR-ready specs; A/C/D have appended CORRECTION sections)
+- `.claude/specs/consumer-crate-template.md` (W8; re-target from hubspo-rs hypothetical to woa-rs/medcare-rs precedent)
+- `.claude/specs/ogit-g-smoke-test.md` (W11 validation)
+- `.claude/specs/trivia-prs-bundle.md` (W12 — 3 quick wins parallel-shippable)
+- `.claude/board/sprint-log-3/{SPRINT_LOG.md,agents/agent-W1..W12.md,meta-1-review.md,sprint-summary.md}`
+
+PR sequence: #360 → #361 → post-#360 substrate-sweep (this PR).
diff --git a/.claude/knowledge/cca2a-sprint-prompt-template.md b/.claude/knowledge/cca2a-sprint-prompt-template.md
@@ -0,0 +1,112 @@
+# CCA2A Sprint Prompt Template
+
+> **READ BEFORE WRITING agent prompts** for any sprint that scopes new architectural work.
+>
+> This template captures the lessons learned from sprint-2 (PR #358/#359) and sprint-3 (PR #360/#361/sprint-3-rescope) about how to prevent the recurring failure modes.
+
+## Failure modes captured to date
+
+1. **"Designing What's Already Built" anti-pattern** (sprint-2 itself; recurred in sprint-3 PR-A-1/PR-C-1/PR-D-1 specs) — agents propose new types/files for substrate that already ships
+2. **Wrong-repo error** (W7 sprint-2 → ndarray; W9 sprint-3 → ada-consciousness) — agents defer to `GITHUB_REPO` env var instead of explicit prompt
+3. **Race-with-merge** (PR #358/#359, PR #360/#361, this PR) — main thread corrections land after PR merge; need follow-up PR
+4. **Pattern letter divergence** (sprint-2 W2 invented A-G that conflicted with W1) — fixed by embedding canonical pattern letters in every prompt
+
+## Substrate-grep checklist (mandatory before any new spec)
+
+Before writing PR-X-Y spec for any pattern X, run:
+
+```python
+import os
+from github import Github, Auth
+tok = os.environ["GITHUB_TOKEN"].strip().strip('"').strip("'")
+g = Github(auth=Auth.Token(tok))
+repo = g.get_repo("AdaWorldAPI/lance-graph")  # CANONICAL — do not defer to GITHUB_REPO env
+
+# 1. Read tier-0 doc — current pattern X status
+tier0 = repo.get_contents(".claude/knowledge/tier-0-pattern-recognition.md", ref="main").decoded_content.decode()
+# Find Pattern X section; note SHIPPED / PARTIALLY SHIPPED / DESIGN PHASE status
+
+# 2. Read cross-source matrix — find Pattern X row
+cross = repo.get_contents(".claude/knowledge/pattern-recognition-cross-source.md", ref="main").decoded_content.decode()
+# Note: shipped substrate file:symbol citations
+
+# 3. Substrate grep — confirm cited substrate exists
+for path in cited_substrate_files:
+    f = repo.get_contents(path, ref="main")
+    # Verify symbols mentioned in tier-0 actually exist in the file
+
+# 4. TD-X check — read corresponding TD row in TECH_DEBT.md
+td = repo.get_contents(".claude/board/TECH_DEBT.md", ref="main").decoded_content.decode()
+# Find TD-X row; status must align with tier-0
+
+# 5. Recent PR check — has this pattern been touched in the last 30 days?
+recent = list(repo.get_commits(since=datetime.now() - timedelta(days=30)))
+# Scan for commits mentioning Pattern X or its TD-X
+```
+
+If any check reveals shipped substrate, the spec MUST cite it and re-scope to "extension of shipped" rather than "construction from scratch".
+
+## Wrong-repo guardrail (mandatory in every agent prompt)
+
+Every worker agent prompt should include this snippet:
+
+```python
+# WRONG-REPO GUARDRAIL: do not defer to GITHUB_REPO env var
+import os
+from github import Github, Auth
+tok = os.environ["GITHUB_TOKEN"].strip().strip('"').strip("'")
+g = Github(auth=Auth.Token(tok))
+repo = g.get_repo("AdaWorldAPI/lance-graph")  # CANONICAL — explicit, not env-driven
+assert repo.full_name == "AdaWorldAPI/lance-graph", \
+    f"WRONG REPO: got {repo.full_name}, expected AdaWorldAPI/lance-graph"
+```
+
+This 8-line block has prevented zero wrong-repo errors so far (because it didn't exist). Going forward, every spawned agent should have it pre-pended to their write protocol.
+
+## Pattern letter discipline
+
+The canonical pattern letter assignment lives in:
+1. `.claude/plans/unified-ogit-architecture-v1.md` (W1 master, sprint-2)
+2. `.claude/patterns.md` Pattern Recognition Framework (W3 sprint-2)
+3. `.claude/knowledge/tier-0-pattern-recognition.md` (post-PR #359 corrected)
+
+Every worker agent prompt MUST embed the canonical 15-letter table inline. Do not invent new letters; do not re-letter substrate inventory items.
+
+If a worker discovers a genuinely new pattern, they MUST surface it as a meta-1-review escalation, not invent letter P / Q / etc. unilaterally.
+
+## Race-with-merge mitigation
+
+When the main thread (or meta agent) catches a defect AFTER the PR opens but BEFORE merge:
+1. Push the fix to the same sprint branch (PR will pick it up on next push)
+2. Comment on the PR with the fix summary so reviewer sees it before merge
+3. If the merge happens before the fix lands: open a follow-up PR off main (this is what PR #359/#361/this PR does)
+
+## Sprint scaffolding pattern
+
+Every sprint should pre-create:
+1. Branch (`claude/<sprint-N>-<scope>`) off main via pygithub
+2. SPRINT_LOG-N.md scaffolding at `.claude/board/sprint-log-N/SPRINT_LOG.md` (worker roster, deliverables table, coordination notes, **canonical pattern letters embedded**)
+3. Worker prompt template (this template)
+
+Then spawn 12+meta agents in parallel; each agent's prompt includes:
+- Sprint context (compact)
+- Their specific deliverable (file path, target size, acceptance criteria)
+- Wrong-repo guardrail snippet
+- Substrate-grep checklist
+- Canonical pattern letter table (inline, not by reference)
+- pygithub-first protocol with quote-stripped GITHUB_TOKEN
+
+## Cross-references
+
+- `.claude/knowledge/tier-0-pattern-recognition.md` — canonical pattern letters + status
+- `.claude/knowledge/pattern-recognition-cross-source.md` — A-O ↔ Pillars ↔ .grok/ ↔ shipped substrate matrix
+- `.claude/board/sprint-log-2/sprint-summary.md` — sprint-2 lessons (W2 invented pattern letters, W7 wrong-repo)
+- `.claude/board/sprint-log-3/sprint-summary.md` — sprint-3 lessons (W9 wrong-repo, recurring; specs over-scoped)
+- `.claude/patterns.md` Pattern Recognition Framework + Anti-Pattern subsection
+
+## Provenance
+
+Captured post-#360 review pass. Reviewer flagged:
+> "Suggested guardrail for the next sprint: bake into every worker prompt a first-action verify step — `git remote -v` (if local) or `mcp__github__get_me` + repo env confirmation before any write. Cost: ~10 LoC per prompt. Saves 1 wrong-repo-rev per sprint."
+
+This template formalizes that guardrail + extends to substrate-grep + pattern-letter discipline.
diff --git a/.claude/knowledge/pattern-recognition-cross-source.md b/.claude/knowledge/pattern-recognition-cross-source.md
@@ -0,0 +1,71 @@
+# Cross-source pattern recognition matrix
+
+> **READ FIRST** alongside `.claude/knowledge/tier-0-pattern-recognition.md`.
+>
+> This workspace has 4 parallel architectural taxonomies that don't currently cross-reference each other. Future sessions reading one and missing the parallels is the recurring failure mode behind the "Designing What's Already Built" anti-pattern. This doc is the canonical mapping.
+
+## The 4 taxonomies
+
+1. **A-O Patterns** (sprint-2 PR #358 + PR #359 + sprint-3 PR #360) — `.claude/plans/unified-ogit-architecture-v1.md`, `.claude/patterns.md`, `.claude/knowledge/tier-0-pattern-recognition.md`
+2. **Pillars 0-4** (palantir-parity-cascade) — `.claude/plans/palantir-parity-cascade-v2.md`, `.claude/plans/ogit-cascade-supabase-callcenter-v1.md`
+3. **`.grok/` epiphanies + open threads** — `.grok/01_overview/` through `.grok/05_query_languages/`, `FANOUT_MAPPING_PLAN.md`, `GLUE_LAYER_OGIT_TO_OWL_SPEC.md`, etc.
+4. **Shipped substrate** (~28 crates) — what actually exists in `crates/`, anchored to TD-X rows in `.claude/board/TECH_DEBT.md`
+
+## Master mapping table
+
+| A-O | Pillar | .grok/ doc | Shipped substrate (file:symbol) | TD ref |
+|---|---|---|---|---|
+| **A** SPO-G u32 slot | Pillar 0 (OGIT u32 namespace) | `04_mathematical_foundation/` | `lance-graph-ontology/src/namespace.rs::SchemaPtr.ontology_context_id` (PARTIAL) | TD-OGIT-G-SLOT-1 |
+| **B** Context Bundle | Pillar 0 (OWL overlay) | `GLUE_LAYER_OGIT_TO_OWL_SPEC.md` | (typed surface missing) | TD-CONTEXT-BUNDLE-2 |
+| **C** Generic Bridge | Pillar 0 (BridgeFromRegistry) | `FANOUT_MAPPING_PLAN.md` § bridge dispatch | `lance-graph-ontology/src/bridges/mod.rs::BridgeFromRegistry` + 3 impls (PARTIAL) | TD-GENERIC-BRIDGE-3 |
+| **D** Meta-Structure Hydration | Pillar 0 (TTL → bundle) | `04_mathematical_foundation/` § hydration | `lance-graph-ontology/src/ttl_parse.rs::parse_ttl_directory_with_provenance` (PARTIAL) | (anatomy demo PR-ANATOMY-1) |
+| **E** Compile-Time Consumer Binding | Pillar 1 (manifest format) | `MCP/` | (manifest format missing) | TD-MANIFEST-MODULES-4 |
+| **F** ractor Supervisor | Pillar 1 (supervised mesh) | `INVESTIGATION_AGENT_CODE_SKETCH.rs` | `cognitive-shader-driver/src/grpc.rs` (shape proven, supervisor missing) | TD-RACTOR-SUPERVISOR-5 |
+| **G** Best-Practice Thinking Inheritance | Pillar 2 (per-G styles) | `03_cognitive_layers/` | `p64-bridge::STYLES` (12-base codebook) | (no TD; Pattern G deferred) |
+| **H** Switchable Cognitive Vessel | Pillar 2 (cognitive-shader-driver) | `01_overview/` | `p64-bridge/src/lib.rs::cognitive_shader::CognitiveShader` ✅ SHIPPED | (no TD) |
+| **I** Implicit Cognition | Pillar 2 (CycleAccumulator) | (parallel to FANOUT) | `lance-graph-contract/src/cycle_accumulator.rs` ✅ SHIPPED | (no TD) |
+| **J** INT4-32D Atoms | Pillar 3 (proximity reranker) | `03_cognitive_layers/` § proximity | `thinking-engine/src/reranker_lens.rs` (lens shape only) | TD-INT4-32D-ATOMS-6 |
+| **K** Circular Compilation | Pillar 3 (JIT actor) | `02_core_primitives/` § JITson | `lance-graph/src/cam_pq/jitson_kernel.rs` (kernel scope) | TD-CIRCULAR-COMPILATION-7 |
+| **L** SPO-Chain Narrative | Pillar 4 (NARS chains) | `05_query_languages/` | `lance-graph-planner::nars::truth::TruthValue` + AriGraph (PARTIAL) | (no TD; deferred) |
+| **M** Wave-Particle Bimodal | Pillar 2 (qualia + SPO) | `03_cognitive_layers/` § wave-particle | bgz17 + qualia (wave) + AriGraph + SPO (particle) ◐ PRIMITIVES SHIPPED | (no TD) |
+| **N** Fingerprint-as-Codebook | Pillar 0 (substrate operation) | `02_core_primitives/` § codebook | `thinking-engine::prime_fingerprint`, `qualia::FAMILY_CENTROIDS`, `p64-bridge::STYLES` ✅ SHIPPED | (no TD) |
+| **O** Phenomenological Memory | Pillar 2 (Qualia17D) | `03_cognitive_layers/` § qualia | `thinking-engine/src/qualia.rs` (39 KB) ✅ SHIPPED | (no TD) |
+
+## Shipped consumer scaffolds (Pattern C precedents)
+
+The Pattern C "consumer template" worked example doesn't need to be hypothetical (hubspo-rs). Two real precedents already validate the LOC reduction claim:
+
+| Consumer | Repo / PR | Scaffold size | Status |
+|---|---|---|---|
+| **woa-rs** | `AdaWorldAPI/woa-rs` PR #2 | ~150 LOC | merged |
+| **medcare-rs medcare-bridge** | `AdaWorldAPI/medcare-rs` PR #110 | ~250 LOC | merged |
+| **smb-office-rs SmbMembraneGate** | `AdaWorldAPI/smb-office-rs` PR #29 | ~30 LOC newtype | merged |
+| **medcare-rs MedCareMembraneGate** | `AdaWorldAPI/medcare-rs` PR #98 | ~30 LOC newtype | merged |
+| (hypothetical) **hubspo-rs** | (W8 spec proposed scaffolding from scratch) | <150 LOC target | PROPOSED, NOT NECESSARY |
+
+The architectural validation gate originally scoped to "build hubspo-rs in <150 LOC" should re-target to "verify woa-rs (~150 LOC) + medcare-bridge (~250 LOC) hit the LOC budget claim" — those are merged precedents.
+
+## Maintenance protocol
+
+When a new architectural insight surfaces in any one of the 4 taxonomies:
+1. **Locate it in this matrix** — find the closest existing row
+2. **Update all 4 columns** — A-O letter, Pillar number, .grok/ section, shipped-substrate file
+3. **Cite the substrate file** with `file:symbol` precision (so future sessions can `grep` to verify before designing)
+4. **If the insight has no existing row** — add a new row only after confirming it's not a re-discovery (use the substrate-grep checklist in `.claude/knowledge/cca2a-sprint-prompt-template.md`)
+
+## Cross-references
+
+- `.claude/knowledge/tier-0-pattern-recognition.md` — canonical A-O letter assignments + status
+- `.claude/plans/palantir-parity-cascade-v2.md` — Pillars 0-4 architecture
+- `.claude/plans/ogit-cascade-supabase-callcenter-v1.md` — Pillars 0-4 cascade execution
+- `.grok/01_overview/` through `.grok/05_query_languages/` — parallel-session epiphanies
+- `.claude/board/TECH_DEBT.md` — TD-X execution backlog
+- `.claude/board/ARCHITECTURE_ENTROPY_LEDGER.md` — per-component scoring + clusters
+- `.claude/knowledge/cca2a-sprint-prompt-template.md` — substrate-grep checklist (next deliverable)
+
+## Provenance
+
+This doc was created post-#360 (sprint-3 specs PR) review pass. Reviewer flagged:
+> "Four parallel architectural taxonomies now exist in this workspace... None of them cross-reference the other three. Future sessions read one and miss the parallels."
+
+This matrix is that cross-reference.