Date · Class · Scope · Limit

Evidence includes what worked, what changed, and what remains open.

A local test establishes software behavior; a simulation establishes a numerical output; a measurement requires an instrument, protocol, and data. Torsión preserves those boundaries.

Open the record ↓

Historical snapshot: August 12, 2026 · focused Verde update: October 5, 2026

Evidence bridges, not absolutes.

August 2026 · F7F/PoTTraceable preflight

Claim–evidence relations remain inactive

Four claim identities and six evidence identities are instantiated across six pairings. Criteria and dossiers exist, while competent reviews and activated relations remain at zero.

Establishes: Identity, provenance, limits, and review readiness.
Does not establish: Physical validation, external acceptance, or empirical evidence.

August 2026 · External reviewPending

The academic route is prepared without implying endorsement

As of the August 2026 snapshot, Torsión was evaluating institutional routes for mathematical and provenance review. At that time, no institution had been selected, contacted, affiliated, or presented as a collaborator or endorser.

Exists: A governed selection, double-analysis, and contact process.
Next step at that snapshot: Human route decision; any contact required later authorization.

October 5, 2026 · Institutional outreachInquiry sent

Orientation for a passive thermal pilot

An orientation inquiry was sent to UAQ to explore measurement capabilities and independent comparison of Verde's thermal model. D9 confirmed sending it; this does not establish delivery, reading, a reply, access to a test bench, institutional collaboration or endorsement.

Scope: Initial inquiry, without a commitment or physical validation.
Missing: Confirm the counterpart, instruments, reference, uncertainty, sampling and person responsible for the bench.

See Verde's technical scope →

July 2026 · Rotor–18Verifiable software

Dimensional correction

The historical version produced 2.84×. CSI-02 v2 corrected the convention to hν = ħω and produces approximately 1.29294 within the computational model. This is not a measured physical effect.

Artifact: Local suite, T1→T2→T3 test, and correction genealogy.
Next step: Citable framework and experimental protocol.

July 2026 · EnergyComputational model

Thermal rule

The 35 °C scenario differs by ~16% from the previous formula. It is not a panel or node measurement.

Artifact: Model, FSM, and local tests.
Next step: Physical pilot with telemetry.

October 5, 2026 · VerdeLocal software · synthetic

Thermal record and passive assembly

53/53 synthetic tests in the focused record-and-assembly suite. They check units, traceability, and software blocks, without physical validation or measured energy efficiency. The agentic matrix is undergoing architectural review and optimization.

Establishes: Local compatibility and reproducibility with synthetic data.
Still needed: Test bench, instruments, an independent thermal reference, uncertainty, sampling, acceptance criteria, and an owner.

July 2026 · OmniPredictVerifiable software

Parameter removed

An underived constant was removed while synthetic integration remained stable. This does not imply clinical validation.

Artifact: M1–M5 pipeline and local suite.
Next step: Derivation and uncertainty.

Update at a verified milestoneGovernance

Corrections, limits, and snapshot date

Public evidence is updated through manual editorial review when a verifiable milestone is available. Each update must preserve source, class, scope, limit and human adjudication; no fixed cadence or automatic generation is promised.

Includes: Provenance, date, scope, corrections, and quarantines.
Does not replace: Measurement, competent review, or authorized publication.