feat: intelligence-loop arc (ADR-0246, 0247, 0248 accepted) #64

Merged
core-labs merged 6 commits from feat/intelligence-loop-arc into main 2026-07-18 18:21:16 +00:00
Contributor

This PR merges the feat/intelligence-loop-arc branch which includes the closing of the 5 seams:

  • S1: Articulation seam closed (core/physics/linguistic_readback.py)
  • S2: Learning write-path activated (evals/analogical_transfer/biography_session.py)
  • S3: Autonomy floor unified (core/physics/cognitive_lifecycle.py)
  • S4: Lift instrument (evals/generalized_lift_instrument.py)
  • S5: Handoff evidence (evals/lift_evidence_handoff.py)

[Verification]: Smoke suite passed locally (~20s, 176 passed)

This PR merges the `feat/intelligence-loop-arc` branch which includes the closing of the 5 seams: - **S1:** Articulation seam closed (`core/physics/linguistic_readback.py`) - **S2:** Learning write-path activated (`evals/analogical_transfer/biography_session.py`) - **S3:** Autonomy floor unified (`core/physics/cognitive_lifecycle.py`) - **S4:** Lift instrument (`evals/generalized_lift_instrument.py`) - **S5:** Handoff evidence (`evals/lift_evidence_handoff.py`) [Verification]: Smoke suite passed locally (~20s, 176 passed)
ai-agent added 6 commits 2026-07-18 18:19:24 +00:00
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.
Egress route readback_eligible now flows into geometric token selection
(resonance = <psi psi~_T>_0 via the I-04 phase correlation; cga_inner is
the un-reversed grade-0 product and is wrong for versor states — pinned in
tests) plus the hearing-ourselves-think round-trip: articulated tokens are
re-ingested through the same sensorium boundary and phase-locked agreement
is measured with the same metric. Fail-closed typed ReadbackRefusal when no
token resonates (decoding, not generating — no fallback strings). Vocab
access is structural (VocabLike) to avoid a core.physics->vocab cycle; the
real VocabManifold satisfies it (proven in tests). 10 new pins, lifecycle
suite untouched (48 passed).
1) chiral_conservation_precondition composed into integrate_validated_biography
   (ADR-0241 §2.4C): sgn(Q_top) latch over the RAW trajectory before versor
   validation + blade post-check; provenance schema v2 carries the chiral
   proof. Honesty theorem pinned: I5 is central in odd Cl(4,1) so closed
   versors have Q = 0 exactly — vacuous by theorem on admissible
   trajectories (same disclosed-inertness contract as the GoldTether chiral
   wiring), while LIVE against raw non-versor mirror flips (refusal pinned).
2) First real caller: evals/analogical_transfer/biography_session.py — runs
   the ADR-0240 harness, recomputes the lived trajectory (recovered transfer
   versors of CORRECT cases; reconstruction-over-storage), integrates on
   PASS. I-01 reboot-invariance pinned BIT-IDENTICAL across runs.
3) biography.py trajectory digest now explicit '<f8' LE bytes (ADR-0245
   §2.3); byte-identical on LE targets, platform-independent everywhere.
31 passed across wiring/session/holonomy/readback suites.
CognitiveLifecycleEngine gains an optional GoldTetherMonitor; solve() feeds
one tether_reading per turn. Control law composed with pin SD-A: closed
certified turns update the monitor (may elevate autonomy), non-admitted /
drifted turns fail-close it (autonomy hard 0), admitted OPEN superpositions
are measured WITHOUT a state update — punishing legitimate interference
states would encode the exact defect egress refuses. Chiral orientation
observed on every reading (material flip raises, fail-closed). Reading is
disclosed on LifecycleOutcome.tether, deliberately outside outcome_id
(observer state is not cognitive content). Residual kernel was already
unified (goldtether delegates to WaveManifold.measure_unitary_residual) —
the audit's 'integrate Monitor into WaveManifold' stays rejected (layering).
49 passed (tether + lifecycle + corridor).
S4: evals/generalized_lift_instrument.py — corridor vs symbolic baseline on
identical compiled problems, independent truth-table gold, honest-NULL
protocol, no-silent-caps (GSM8K non-ingestibility RECORDED as the
composition frontier). Live results: propositional PARITY (corridor 10/10 =
ROBDD 10/10, wrong=0 guard HELD), constrained-recognition LIFT +9 (relax+
readback 9/9 vs constraint-blind argmax 0/9; round-trip agreement 1.0),
multimodal-completion PARITY (vision token already resonates with the
audio partial — measured, disclosed).

S5: evals/lift_evidence_handoff.py — two REAL certified turns through
IdentityPort+PrecisionPort (Ring-2) and coordinate_handoff (Ring-3),
off-serving, flags untouched: identity-action turn PROCEEDs, frame-rotating
turn ABSTAINs with typed port-attributed reason (d_stab>epsilon_turn); both
replay chains verify. Evidence doc: docs/handoff/ADR-0246-Acceptance-
Evidence.md (for Shay's §8 rulings — no self-Accept). Plan doc phases 0-5
marked DONE.

[Verification]: Smoke suite passed locally (129s, 176 passed); 147 passed
across all touched+adjacent suites.
Approved via explicit human ratification.
core-labs merged commit 31f1a824b9 into main 2026-07-18 18:21:16 +00:00
core-labs deleted branch feat/intelligence-loop-arc 2026-07-18 18:21:21 +00:00
Sign in to join this conversation.
No reviewers
No labels
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference: core-labs/core#64
No description provided.