docs(research): intelligence-loop arc opening record — plan + Spark-audit adjudication

Claim-by-claim adjudication of the two Google Spark audit PDFs against
main @ 54659b76 (most claims closed by the 2026-07-17 waves; two
recommendations rejected as doctrine violations), plus the phased
homestretch plan (S1-S3 seams, generalized-lift instrument, ADR-0246/47/48
acceptance evidence). Source PDFs included for provenance.

docs/research/intelligence-loop-homestretch-plan-2026-07-18.md

@@ -0,0 +1,114 @@
+# Intelligence-Loop Homestretch Plan
+
+**Status**: Proposed plan — awaiting Shay's go before any implementation
+**Date**: 2026-07-18
+**Base**: `main @ 54659b76` (Rings 2+3 merged; ADR-0243/0244/0245 Accepted; ADR-0246/0247/0248 Proposed, flags off)
+**Foundation**: `docs/research/spark-audit-adjudication-2026-07-18.md` — every work item below survived
+claim-by-claim verification against the live tree; stale audit claims were removed there and must not
+resurface as work items.
+**Related**: ADR-0240, ADR-0243…0248, `docs/research/core_cohesion_master_plan.md`
+
+---
+
+## 1. Objective (done-when)
+
+Close the comprehend → reason → articulate → contemplate → learn loop and make its value measurable:
+
+1. **All green CI** (smoke + fast lanes) at every phase boundary.
+2. **Meaningful, generalized lift in problem solving, without overfitting** — decided by a dedicated
+   instrument (Phase 4), not by narrative. An honest NULL is a legitimate, recordable outcome.
+3. **Articulate comprehension of inputs and articulate generation of outputs** as one cumulative loop —
+   the corridor's field→token readback wired and round-trip-verified (Phase 1).
+4. `generate/` ("core_logos" role) upgrade explicitly **deferred** until the above land (Phase 6 scope note).
+
+## 2. Governance constraints binding all phases
+
+- **A-04 serve-path containment**: wave-field / Fibonacci operators stay out of `chat/runtime.py`;
+  corridor work lives in `core/physics/` + `evals/` quarantine. Serving consumes only ratified frozen artifacts.
+- **I-03 / proposal-only learning**: no autonomous mutation of the active manifold; biography integration
+  only via the PASS-gated wiring layer, harness-driven.
+- **No self-Accept**: ADR-0246/0247/0248 stay Proposed; phases produce acceptance-packet evidence only.
+  Flag flips and Accepts are Shay's alone.
+- **Decoding, not generating**: readback refusals are typed and fail-closed; no fallback strings.
+- **Worktree discipline**: all work in a worktree off `forgejo/main`; one PR per coherent phase;
+  smoke gate in-worktree before any push.
+
+## 3. Phases
+
+### Phase 0 — Worktree + record (small) — PARTIALLY DONE
+- [x] Spark PDFs copied to `docs/research/`.
+- [x] Adjudication doc written (`spark-audit-adjudication-2026-07-18.md`).
+- [x] This plan doc written.
+- [ ] Create worktree off `forgejo/main`; run smoke + fast lanes → green baseline recorded.
+- [ ] Commit the three docs + PDFs as the arc's opening record.
+
+### Phase 1 — Close the articulation seam, S1 (medium)
+Wire a `readback` stage into the lifecycle corridor (eval-tier, off-serving):
+- `egress` `route="readback_eligible"` → `VocabManifold.nearest()` token selection
+  (`cognitive_lifecycle.py` gains a vocab consumer; today it imports no `vocab`, `:68-76`).
+- Fail-closed: typed refusal when no resonant token exists — no fallback strings.
+- **Round-trip feedback eval ("hearing ourselves think")**: re-ingest the articulated token sequence
+  and measure phase-integrity / resonance agreement against the pre-articulation steady state.
+  Falsifiable pass criterion fixed before implementation.
+- TDD anchors: `tests/test_adr_0243_cognitive_lifecycle.py`, `tests/test_vocab_manifold_invariants.py`.
+- Scope guard: within Accepted ADR-0243 §2.3 (readback rules); if design exceeds it → ADR amendment, not silent drift.
+
+### Phase 2 — Activate the learning write-path, S2 (medium)
+- Compose the chiral Q_top latch (`chiral_gate.py`) into `integrate_validated_biography`
+  (`biography_wiring.py:174`): charge conservation becomes a precondition of biography updates.
+- First real caller: harness-driven — after ADR-0240 validation PASS, integrate holonomy; prove
+  **I-01 reboot-invariance** and replay determinism in tests.
+- Strictly PASS-gated + harness-driven; no idle-loop auto-integration (I-03).
+- Cleanup-as-you-find: `biography.py:62` bare `.tobytes()` → explicit `<f8` coercion.
+- TDD anchors: `tests/test_adr_0240_biography_holonomy.py`, `tests/test_adr_0243_biography_wiring.py`.
+
+### Phase 3 — Unify the autonomy floor, S3 (small)
+- Route the lifecycle's residual check (`cognitive_lifecycle.py:831`) through `GoldTetherMonitor`
+  so autonomy-level modulation + chiral latching govern corridor runs — one guard, not two half-guards.
+- Layering direction: lifecycle → goldtether → WaveManifold (the audit's "integrate Monitor into
+  WaveManifold" is rejected; see adjudication §2).
+- Not wired into serve (A-04). Parallelizable with Phases 1–2.
+
+### Phase 4 — Generalized-lift instrument (medium-large, the crux)
+Instrument before consumption (ADR-0190 lesson: build the measurement first):
+- Corridor-vs-symbolic-baseline evaluation harness: problem Hamiltonians compiled from the same
+  readers both paths use; measure solve success + refusal honesty.
+- Panels: 3-domain anti-overfit panel; deductive flagship as a **guard** (wrong=0 must hold);
+  real GSM8K holdout slice (never the 150 templated cases); field-reasoner reading-independence wedge.
+- Anti-overfit discipline: holdout sealed; no per-case tuning; κ/γ calibration only through the
+  existing sealed calibration evals + ratification.
+- Honest-NULL protocol: if the corridor adds no eval delta, record NULL (as ADR-0246 §11 did).
+  Truth test is eval delta, not artifact append.
+
+### Phase 5 — Handoff evidence for ruling (medium)
+- Run the Phase 4 instrument through ADR-0247 ports + ADR-0248 handoff machinery (flags on in
+  evals only) and assemble acceptance-packet evidence for ADR-0246/0247/0248.
+- Deliverable: packets ready for Shay's §8 rulings. No self-Accept, no flag flips.
+
+### Phase 6 — Deferred: `generate/` ("core_logos") upgrade
+- Out of scope for this arc by explicit direction. Phase 4's results show where articulation
+  quality actually binds; that evidence scopes the next arc.
+
+## 4. Dependencies
+
+- Phase 1 → Phase 4 (instrument needs the closed loop). Phases 2–3 parallel with 1.
+- Phase 5 needs Phase 4's evidence. Each phase = own PR, smoke-gated, merged before the next builds on it
+  (stacked only if necessary; merge-commit discipline).
+
+## 5. Risks (honest register)
+
+| Risk | Level | Handling |
+| :--- | :--- | :--- |
+| Phase 4 returns NULL lift (corridor doesn't beat symbolic path yet) | **High-likelihood, low-harm** | Honest NULL recorded; instrument makes the question decidable; scopes Phase 6 |
+| VocabManifold coverage too sparse for meaningful readback | Medium | Fail-closed refusals are correct behavior; coverage growth is proposal-gated corpus work, not this arc |
+| New eval suites bloat fast lane / flake in CI | Medium | Lane markers (`slow`/`quarantine`) assigned at authoring time; smoke gate at every phase |
+| Biography wiring drifts toward autonomous mutation | Low | Harness-driven, PASS-gated only; I-03 checked in review at Phase 2 PR |
+| Scope creep beyond Accepted ADR text | Low | Each phase names its governing ADR section; exceeding it → ADR amendment PR, not silent drift |
+
+## 6. Decisions to make at phase start (not blockers now)
+
+- Phase 1: exact round-trip pass criterion (resonance-agreement threshold vs rank-agreement of top-k tokens).
+- Phase 2: whether the first caller lives in `evals/analogical_transfer/harness.py` or
+  `evals/adr_0243_cognitive_lifecycle/` (leaning: adr_0243 harness — it already imports GoldTetherMonitor).
+- Phase 4: GSM8K holdout slice size + which 3 panel domains (reuse the universal-structure trio unless
+  evidence says otherwise).

docs/research/spark-audit-adjudication-2026-07-18.md

@@ -0,0 +1,95 @@
+# Spark Audit Adjudication — claim-by-claim verification against `main`
+
+**Status**: Record (verification complete 2026-07-18)
+**Verified against**: `main @ 54659b76` (post Rings 2+3 merge)
+**Inputs**: `docs/research/core_labs_core_technical_gap_audit.pdf` ("Evolutionary Landscape", 8pp),
+`docs/research/core_repository_systems_audit_and_gap_analysis.pdf` (systems-grade review, 4pp)
+**Method**: three independent read-only verification passes over the live tree; every claim
+adjudicated with file:line evidence before being allowed to become a work item.
+**Companion**: `docs/research/intelligence-loop-homestretch-plan-2026-07-18.md` (the plan built on this record)
+
+---
+
+## 1. Headline finding: the audits are one day stale
+
+Both documents describe the repository **before the 2026-07-17 cohesion-directive waves**
+(D1/D2, ADR-0244 Phases 2–6, ADR-0243 flagship lifecycle, ADR-0246 rings). Most named gaps
+were closed on `main` the day before the audits were delivered:
+
+| Closing commit | What it closed |
+| :--- | :--- |
+| `c3b4d045` feat(adr-0243) flagship cognitive_lifecycle.py | ingress → relax → egress orchestration |
+| `4ec2c8b9` refactor(adr-0244) D1 semantic rigor | full 256-bit digests + LE byte-order |
+| `82b031d1` perf(adr-0244) D2 mechanical sympathy | Rust f64 GP fast-path + `_cached_eigh` memoization |
+| `5c69b741` Phase 5a §2.7 content-id semantic rigor | remaining content-id sites |
+| `918aa843` Phase 4 governed f64→f32 serving-boundary cast | Serves-Boundary Cast Contract |
+| `5027adf8` ADR-0244 + ADR-0245 Proposed→Accepted (Joshua Shay) | audits assume both still Proposed |
+
+Rule reaffirmed: **external audit claims are point-in-time and must be verified against source
+before being acted on** (see also `feedback-read-source-not-sandbox-traces`).
+
+---
+
+## 2. Claim-by-claim verdicts
+
+### Refuted / already closed
+
+| # | Audit claim | Verdict | Evidence |
+| :--- | :--- | :--- | :--- |
+| 1 | "Relax phase omitted — orchestrator skips ingest→egress" | **REFUTED** | `CognitiveLifecycleEngine.solve()` runs `ingest_context` → `relax_to_ground` → `egress` (`core/physics/cognitive_lifecycle.py:1054-1071`); `relax_to_ground` (`:530`) is deterministic imaginary-time relaxation with a certified `RelaxationCertificate` |
+| 2 | "Hardcoded `[RESONANT_LOCK…]` mock readback; sine-wave proxies" | **REFUTED** | No such string anywhere (grep empty); real geometric readback exists: `VocabManifold.nearest()` = exact CGA inner-product argmax, "no cosine similarity, no ANN index" (`vocab/manifold.py:261-298`) |
+| 3 | "`_cached_eigh` LRU completely missing in identity.py" | **REFUTED** (category error) | ADR-0245 §2.4 targets the frozen ProblemHamiltonian, not identity.py; implemented exactly to spec at `cognitive_lifecycle.py:509-522`, keyed `(hamiltonian_id, H_mat.tobytes())`, called at `:588`. `identity.py` performs zero eigendecompositions. ADR-0245 status map records §2.4 "✅ Done" |
+| 4 | "Vault digests truncated to 24 hex chars" | **REFUTED** | Vault hashing is full 64-hex: `vault/store.py:513-515`, `vault/crdt.py:204`, `core/physics/holographic_vault.py:80`. Only human-readable labels truncate (`core/demos/*`, `core/cli_teaching.py:626,977`) — explicitly permitted by ADR-0245 §2.3 |
+| 5 | "Platform-dependent hashing (no `<f8` coercion)" | **REFUTED** for all load-bearing sites | `_le_f64_bytes` (`cognitive_lifecycle.py:137-138`), `holographic_vault.py:79`, `identity_action.py:223,528` — all coerce `np.dtype("<f8")`. Lone hygiene residual: `biography.py:62` hashes bare `float64.tobytes()` (byte-identical on LE targets, outside the vault content-address set) — folded into plan Phase 2 |
+| 6 | "Fibonacci API drift (missing `best_observed_point`/`eval_sequence`)" | **REFUTED** | `SearchTrace` compat dataclass carries exactly those fields (`fibonacci_search.py:108-112`) via `search_trace_from_result` (`:368`); `FibonacciSearchCertificate` keeps `minimizer`/`ordered_points`/`ordered_values` (`:60-63`) |
+| 7 | "GoldTether residual math duplicated inline elsewhere" | **REFUTED** | One shared kernel: `goldtether.py:117-125` `coherence_residual` delegates to `WaveManifold.measure_unitary_residual`; `cognitive_lifecycle.py:831` calls the same kernel |
+| 8 | "Biography holonomy not implemented — no lifelong learning mechanism" | **REFUTED as stated** (see §4 for what IS open) | `integrate_biography` builds the holonomy blade via `holonomy_encode` (`core/physics/biography.py:66-110`); pinned by `tests/test_adr_0240_biography_holonomy.py:17-51`; PASS-gated wiring layer exists (`biography_wiring.py:174`) |
+
+### Rejected as doctrine violations — must NOT be applied
+
+| Audit recommendation | Why rejected |
+| :--- | :--- |
+| "Wire Fibonacci search to run online calibration on live sensory streams" | Violates master-plan **A-04** (serve-path containment: wave/Fibonacci operators quarantined to `evals/` + calibration), **I-03** (no self-mutation), **R-04** (hot paths receive only pre-ratified frozen scalars). Calibration correctly lives in `evals/adr_0244_gamma_calibration/__init__.py:187` and `evals/analogical_transfer/kappa_calibration.py:30` (via `goldtether.propose_kappa_line_search`, `goldtether.py:567-606`) |
+| "Integrate GoldTetherMonitor directly into WaveManifold" | Inverts layering: `goldtether.py` imports `WaveManifold`, not vice versa. Correct fix is the opposite direction (plan Phase 3) |
+
+---
+
+## 3. Naming / provenance reconciliation
+
+- The audits' `multimodal_lifecycle.py` = the real `core/physics/cognitive_lifecycle.py`.
+- The audits' `vocab/readback_rules.py` = the real `vocab/manifold.py` (`VocabManifold`).
+- **`core_logos` does not exist as code.** It was the predecessor `core-ai` subsystem;
+  `docs/adr/ADR-0011-renderer.md:96` states no `core_logos` equivalent will be introduced.
+  The serve-time articulation role is played by `generate/` (`realizer.py`, `articulation.py`,
+  `surface.py`, `stream.py`), consumed by `chat/runtime.py` (`:95-111`). "core_logos is the
+  weakest link" therefore maps to the `generate/` stack.
+- The `docs/research/` ADR-0241…0243 documents are the Spark research series that seeded the
+  canonical `docs/adr/` ADR-0241…0245 (now **Accepted**, ratified 2026-07-17). The canonical
+  ADRs supersede the research drafts wherever they diverge.
+
+---
+
+## 4. Confirmed real seams (survived verification — these ARE the work)
+
+- **S1 — Articulation seam.** The lifecycle's `egress_gate` stops at `route="readback_eligible"`
+  (`cognitive_lifecycle.py:803,852`) and never invokes `VocabManifold.nearest()`
+  (`cognitive_lifecycle.py` imports no `vocab`, `:68-76`). The field→token step is unwired even
+  inside the quarantine corridor; the comprehend→reason→articulate loop does not close, and the
+  "hearing ourselves think" feedback (re-ingesting own articulation) does not exist.
+- **S2 — Learning write-path dormant.** `integrate_validated_biography`
+  (`biography_wiring.py:174`, eval-report-PASS gated) has zero live callers (tests only).
+  Topological-charge conservation exists (`chiral_gate.py:1-114`, sgn(Q_top) latching) but is
+  wired into `GoldTetherMonitor` (`goldtether.py:160,242`), decoupled from biography updates
+  (which use the wave unitary residual instead). Machinery built; nothing drives it; the two
+  guards are not composed.
+- **S3 — Autonomy floor bypassed in the corridor.** `cognitive_lifecycle.py:831` checks the raw
+  unitary residual directly, bypassing `GoldTetherMonitor` autonomy-level modulation + chiral
+  latching (only `self_authorship.py:15` consumes the Monitor). One guard should govern, in the
+  goldtether→WaveManifold layering direction.
+- **S4 — The serving bridge is the ratification frontier.** Serving is the symbolic `generate/`
+  path; the wave corridor is Tier-2 off-serving by design (`core/physics/__init__.py:156,187`
+  TYPE_CHECKING-only; runtime instantiation only in `evals/adr_0243_cognitive_lifecycle/__init__.py:117,119`
+  and tests). The governed seam machinery — ADR-0247 ports + ADR-0248 handoffs — is merged,
+  flag-off, **Proposed**, awaiting evidence and Shay's ruling.
+- **S5 — `generate/` upgrade deferred** by explicit direction until S1–S3 + the measurement
+  instrument land (plan Phase 6).