Kessler · Lattice OS
Kessler Audit Charter
Per-claim verification of the Kessler methodology against the public DISCOSweb catalog. Each row below corresponds to a downloadable JSON audit artifact with claim ID, tolerance, pinned inputs, and result. Every figure on this page is read from those JSONs.
What this page renders. Audit-claim results computed by the Kessler compiler against the public DISCOSweb catalog. The compiler implementation (substrate algebra, leverage-map computation, Klein-4 embedding) is private IP and not published.
How this edition was run. Every check was executed again by its own validator or generator on 2026-11-05, in dry-run mode: the daily compile wrote its lookups into an in-memory stand-in for the production database, and the customer API ran in-process on those lookups. No production database, sign-in service, or deployed endpoint was read or written. The DISCOSweb catalog was not pulled again for this edition (the pull needs an account token that this run did not use), so every catalog-based figure rests on the leverage map of 2026-04-28, SHA-256 47f32fd6942384e282c2691bb4f8a0142be07f4843dfb10fc27dfcd7027cdd40. The GFZ geomagnetic and solar-flux archive was refreshed to 2026-11-04. Rows marked PENDING could not be run without account data; they are listed under Not re-verified in this edition.
What is published. The underlying mathematics is published as a family of Zenodo preprints with associated provisional patents (listed below). The full Kessler Critical Surface preprint is the synthesis artifact for this audit charter, forthcoming. Each claim below verifies a specific testable consequence of the published mathematics.
Engagement scope. Per-customer fleet analysis (substrate-coupling pathway extraction with named-target attribution, atomic intervention prescriptions, refreshable subscriptions) is delivered only on engagement. tony@fancyland.net.
What changed since v0.4.1
- Every check re-executed. Of the rows whose v0.4.1 JSON exists, 25 of 32 show the same headline value in this edition. Recomputed from the v0.4.1 inputs (its archive, forecast start and cohort parameters), the v0.4.1 MTBOF, M6 and M2 values reproduce bit for bit: 30 of 30 (M1(b)). The leverage portfolio, the intervention, the coupling pathway and the T-state cohort, which had no generator in the repository, now have one; each reproduces its v0.4.1 artifact exactly.
- Checks added. M1 PASS, M4 FAIL, M5 NOT_IMPLEMENTED, A4 FAIL (all four PENDING in v0.4.1); F2b PASS (the archive refresh changes only values GFZ marks preliminary) and Op2 FAIL (every lookup the API reads is written by the daily compile) are new. A13 now re-derives its analytic derivative against finite differences, and T1 to T3 are checked by independent quadrature and against the leverage map. C3, C4 and C6 must now hold from every one of several independent Lanczos starts, and C7a requires the input file to be the pinned one.
- Statuses now read from their JSONs. M3: the v0.4.1 charter showed PASS while its own JSON recorded FAIL; the 800 km canonical orbit reaches the 100-year integration cap under all three archetypes, so its envelope has no spread. D1 was shown as CLOSED VIA M2 over a JSON verdict of FAIL. M2 and A8 were shown as PASS_PARTIAL and CLOSED VIA M2 with 0 events inside the chartered tolerance. S2 and S3 carry the warning their JSONs record. Each is now shown as computed.
- A17 read as written. v0.4.1 described βrel as strictly monotonic over the well-populated shells; it falls from the 250–300 km shell to the next one and rises strictly only over 300–450 km. The analyzer's own rule (more steps rise than fall) still passes; the strict reading is FAIL.
- F3 computed for the first time. The conjunction-graph Fiedler value and IPR were computed by the daily pipeline's own code and written to the public aggregate. The published IPR, 202.6, is the participation ratio and lies outside the range the invariant expects; its reciprocal, 0.004936, would lie inside. F3 is FAIL until the definition is settled.
- Figures that moved. The MTBOF envelope now starts from 2026-11-05; at 600 km its P10 / P50 / P90 moved from 29.65 / 13.80 / 9.71 to 29.62 / 13.44 / 9.80 years. M6 now finds 3 of 3 documented reentries inside the envelope (1 in v0.4.1): the compiler's historical cohort now carries Skylab A/m 0.0024 (was 0.0036) and UARS A/m 0.0120 (was 0.0086) m²/kg, a change made in the code after v0.4.1 was published. The F2 storm-frequency table now spans 34,642 days.
- Figures withdrawn or corrected. A13's v0.4.1 figure “42 quality fits, R²=0.967” has no source in any v0.4.1 JSON and is not carried forward. The intervention subset's altitude now comes from the pinned map (601–619 km mean altitude); the v0.4.1 text gave a fixed figure the map does not support. The intervention baseline agrees with C7a only as its tolerance states, equal when both are rounded to 6 decimal places (the precision the charter shows), not to the last digit.
- Public surface. Every JSON cites this charter at kessler.latticeos.org, never a retired host or a repository. The fleet-member scrub now also covers named members of the other active commercial fleets in the slot lists. Pinned inputs check out with the same bytes on every platform.
Substrate-level claims C1 – C8
Load-bearing for the customer-facing modules. C8 gates whether the substrate framing is empirically grounded.
| ID | Claim | Tolerance | Result | v0.4.1 | JSON |
|---|---|---|---|---|---|
| C1 | Substrate size: φ(510510) coprime residues on (ℤ/m)× | exact (integer) | PASS 92,160 | PASS | JSON ↓ |
| C2 | τ_perm, σ_perm are involutions and commute as permutations | exact (integer) | PASS τ fixed points 64 | PASS | JSON ↓ |
| C3 | V₄ invariance via Frobenius leakage η(m); σ₄/σ₀ at the one-ninth overtone target; parity-chiral degeneracy. Must hold from every Lanczos start the JSON records. | eta <= 1e-13 AND |sigma4/sigma0 - 1/9| <= 1e-6 AND parity-chiral split <= 1e-9 | PASS η = 1.824e-15; largest over starts 1.824e-15 | PASS | JSON ↓ |
| C4 | V₄ σ-invariance: eigenvalues of the projected σ operator are ±1; four-pair qubit basis. Must hold from each start. | abs(eigval) - 1 <= 1e-6 | PASS max residual over starts 2.22e-15 | PASS | JSON ↓ |
| C5 | Klein-4 embedding consistency on the pinned catalog | exact (integer count) | PASS σ 46,492/46,492; τ 49,998/49,998 | PASS | JSON ↓ |
| C6 | σ_mag is constant on Klein-4 orbits (23,024 clean orbits). Must hold from each start. | max within-orbit spread ≤ 1e-12 | PASS max spread over starts 6.91e-15 | PASS | JSON ↓ |
| C7a | Frozen-catalog reproducibility: the pinned leverage map (SHA-256 47f32fd6…) gives ⟨P_σ⟩_LEO = +0.2410. Bit-identical to v0.4.1: True. | input SHA-256 equals the v0.4.1 pin AND |measured - 0.2410| <= 1e-3 | PASS +0.241036 (dev 3.6e-05) | PASS | JSON ↓ |
| C8 | Orthogonal control: ⟨P_σ⟩_LEO over the classical active/debris ratio. Distinguishes substrate-algebraic structure from relabeling. | |ratio - 1| >= 0.30 -> SUBSTRATE_REAL_WORK; <= 0.10 -> MOSTLY_CLASSICAL; between -> MIXED_SIGNAL | PASS ratio = 1.970 → SUBSTRATE_REAL_WORK | PASS | JSON ↓ |
Slot module claims S1 – S7
The Slot Assessment. In this edition the lookups are the ones the daily compile wrote for this snapshot, and the API ran in-process on them; the deployed endpoint was not called.
| ID | Claim | Tolerance | Result | v0.4.1 | JSON |
|---|---|---|---|---|---|
| S1 | Single source of truth: the API reads envelope_definitions from the snapshot, no parallel constants. Canonical orbits classify; edge cases fall back without error. | len(sync_gaps) == 0 AND all canonical orbits classify exactly AND all edge cases return non-error envelope | PASS 6/6 canonical; 0 sync gaps | PASS | JSON ↓ |
| S2 | σ_mag(inc) bins reproduce from the catalog; within-bin std reported alongside the mean | PASS = reproducibility (max bin_mean diff <= 1e-12); WARN if median within-bin std > 0.3 | PASS_WITH_WARN 238/238 bins exact; median std 0.345 | PASS | JSON ↓ |
| S3 | ρ(inc, alt) cells reproduce from the catalog; within-cell relative std reported | PASS = reproducibility (max cell_mean diff <= 1e-6); WARN if frac of well-sampled (n >= 10) cells with rel_std > 0.3 exceeds 30% | PASS_WITH_WARN 2,894/2,894 cells exact | PASS | JSON ↓ |
| S4 | SLS arithmetic through the /v1/slot/assess route: BMD-style envelopes, cascade-trigger list, snapshot pin; missing token, invalid token and bad input refused. Run in-process; the deployed endpoint was not called. | exact integer match on SLS components; envelope name equality; snapshot_id == 2026-11-05 | PASS 4/4 cases | PASS | JSON ↓ |
| S5 | Recommended alternatives: top-K within-envelope by lowest |σ_mag| match the leverage-map ranking | top-K set equality per envelope; order match informational | PASS 6/6 envelopes | PASS | JSON ↓ |
| S6 | Cascade-trigger proximity: top-K named cascade objects per shell ranked by D_mom × |L_signed|; cosparId-verifiable, deterministic | shell classification exact; named targets verifiable via cosparId; |L_signed| ranking descending; deterministic re-run | PASS 5/5 queries; 23/23 cosparIds resolved | PASS | JSON ↓ |
| S7 | Safe-harbor proximity: named substrate-neutral residues (lowest |σ_mag|) within the shell, ranked by operational proximity, with counterfactual SLS at co-location. The lookup is built by the compiler's extractor (see Op2); the route ran in-process. | lookup sorted by |sigma| asc; cosparId 100% verifiable; the slot route returns harbors; counterfactual SLS arithmetic exact; deterministic re-runs | PASS 4/4 queries; 371 of 372 harbors resolved, 1 without cosparId | PASS | JSON ↓ |
Flare module claims F1 – F3
| ID | Claim | Tolerance | Result | v0.4.1 | JSON |
|---|---|---|---|---|---|
| F1 | Loss-response tensor calibrated against regime-spanning historical events: Halloween 2003, Gannon May 2024, Starlink Feb 2022 | per-event tolerance defined in flare_assessor.CALIBRATION_TOLERANCE; the calibration gate IS the methodology PASS criterion | PASS 3/3 events | PASS | JSON ↓ |
| F2 | GFZ Kp/ap archive storm-frequency table, recomputed on the refreshed archive | coverage >= 90 yr; non-negative band counts; below-storm (Kp<5) majority of days; storm-class (Kp>=5) within [5%, 30%]; G5 < 0.5%; computation invariants only -- specific frequencies drift as archive updates | PASS 1932–2026; 34,642 days | PASS | JSON ↓ |
| F2b | Archive continuity: the refresh changes only values GFZ marks preliminary (its own definitive flags, read from the previous edition's source file); the parser reproduces the previous archive byte for byte | parser reproduces the previous archive byte for byte; 0 changed days among those the previous source flagged D=2; 0 changed Kp/ap values among D=1 days; no day missing; D=0 days and F10.7 values may change (reported) | PASS 0 of 34,334 definitive days changed; 27 preliminary days revised | new | JSON ↓ |
| F3 | λ₂ + IPR of the conjunction graph computed against the current snapshot and written to the public aggregate (invariants only) | lambda_2 > 0; IPR in (0, 1]; n_components >= 1; n_objects > 0 | FAIL λ₂ = 1.0123e-05; ipr_v2 = 202.6 | PASS_INFO | JSON ↓ |
Disposal module claims D1 – D3
| ID | Claim | Tolerance | Result | v0.4.1 | JSON |
|---|---|---|---|---|---|
| D1 | King-Hele propagator + Vallado atmosphere with period-appropriate F10.7 and the bundled SWADI correction, against Skylab, UARS and ROSAT. Superseded by the M-class time machine (decision #29). | |predicted − observed| / observed ≤ 0.50 | FAIL 1 of 3 within | CLOSED VIA M2 | JSON ↓ |
| D2 | SLH integrates per-orbit σ-exposure: non-negative, monotone, |σ_mag| linear, σ-symmetric | non-negative; monotone; |sigma| linear within 1e-9; symmetric within 1e-9 | PASS scaling 2.000000; symmetry diff 0.0e+00 | PASS | JSON ↓ |
| D3 | Disposal-baseline percentiles match the cohort-sampled SLH distribution. The daily compile writes the lookup with no cohort yet. | exact (sorted-index lookup) | PASS_INFORMATIONAL 0 samples in the lookup | PASS_INFO | JSON ↓ |
A-class — Atmospheric measurement A1 – A17
SWADI sits between substrate and disposal: substrate validates → SWADI measures atmosphere using substrate-weighted instruments → disposal/MTBOF predicts using SWADI. The SWADI and storm-response inputs are the bundled output of the 2026-04-28 pipeline run, which used space-track.org TLE history; they are not recomputed here, and the verdicts below are recomputed from the bundled statistics.
| ID | Claim | Tolerance | Result | v0.4.1 | JSON |
|---|---|---|---|---|---|
| A1 | Cohort population reproducibility from a SHA-pinned TLE pull | exact cosparId set match | PENDING needs the TLE history | PENDING | — |
| A2 | Per-object dh/dt linear-fit reproducibility; orbit-propagator cross-check | dh/dt difference within the chartered bound | PENDING needs the TLE history | PENDING | — |
| A3 | Maneuver detection: ISS, Hubble and an active mid-LEO mega-constellation member classify correctly | ISS = maneuver; Hubble noisy or clean | PENDING needs the TLE history | PENDING | — |
| A4 | Substrate-weighted F factor bounds on the populated shells; no clamping fires (bundled cohort) | fraction of populated shells with F in [0.3, 3.0] >= 0.99 AND no shell clamped | FAIL 8 of 9 shells in band; the 800–900 km shell reads 3.61; 0 clamped | PENDING | JSON ↓ |
| A5 | Substrate weighting reduces F-factor variance against unweighted, every shell | substrate-weighted below unweighted, all shells | PENDING needs per-object F samples | PENDING | — |
| A6 | Diurnal-bulge signature: SSO against retrograde F differ at mid-LEO | relative difference above the chartered bound | PENDING needs per-inclination F samples | PENDING | — |
| A7 | Held-out disposal cohort calibration | mean relative error within the chartered bound | PENDING needs a locked held-out cohort | PENDING | — |
| A8 | D1 closure through the M-class time machine (the M2.a real-ap test) | M2.a per event: |rel_err| <= 0.20; M2.b and M2.c are controls | FAIL 0 of 3 events within | CLOSED VIA M2 | JSON ↓ |
| A13 | King-Hele pseudomodel: the analytic dh/dt(h) matches the finite difference of h(t); the fitter recovers known parameters. Real Ghost Fleet fits are not re-verified (TLE history). | finite-difference rel-err <= 1e-5 at every sample (as chartered); recovered A, H_eff and t_reentry within 1e-6 relative on every synthetic case | PASS max finite-difference rel-err 5.3e-06 | PASS | JSON ↓ |
| A15 | Per-object normalized-enhancement against ap_daily correlation; storm-response sign per object | median r > +0.30 AND >= 70% of objects positive (as chartered) | PASS_PARTIAL 81% positive; median r +0.19 | PASS_PARTIAL | JSON ↓ |
| A16 | LOAD-BEARING. Cross-object pairwise residual correlation in the same shell and calendar window. Strongest co-orbital active-payload pair r = +0.949 across 159 days. | median r > +0.40 AND >= 70% of pairs positive (as chartered) | PASS_PARTIAL 72% positive; median r +0.12 | PASS_PARTIAL | JSON ↓ |
| A17 | Per-shell drag-enhancement coefficient βrel monotonic in altitude over the well-populated shells; values in the published CHAMP/GRACE band | monotonic across the well-populated shells (as chartered), read strictly; band: every shell used for monotonicity has beta_rel in [0.005, 0.018] | FAIL strict rise only over 300–450 km; band PASS | PASS | JSON ↓ |
M-class — MTBOF Lifecycle Service M1 – M6
Decision #29: MTBOF (Mean Time Between Orbital Failures) replaces the disposal certification. The envelope replays three solar-cycle archetypes forward from the forecast start.
| ID | Claim | Tolerance | Result | v0.4.1 | JSON |
|---|---|---|---|---|---|
| M1 | Time-machine reproducibility: (a) every M3 value recomputed by a second call; (b) the v0.4.1 MTBOF, M6 and M2 values recomputed from the v0.4.1 inputs | |run1 - run2| <= 1 day for every value, in both parts | PASS (a) 12/12; (b) 30/30 bit-identical | PENDING | JSON ↓ |
| M2 | Historical D1 closure through the M2.a / M2.b / M2.c gauntlet against Skylab, UARS and ROSAT. M2.a real-ap rel_err: Skylab -54.2%, UARS +56.5%, ROSAT -30.1%. The methodology ships as a forward-prediction envelope, not a historical-calibration tool (decision #29). | M2.a per event: |rel_err| <= 0.20; M2.b and M2.c are controls | FAIL 0 of 3 within | PASS_PARTIAL | JSON ↓ |
| M3 | P10 / P50 / P90 envelope monotonic with P10/P90 above the chartered ratio on every canonical disposal orbit. 600 km, A/m 0.01 → 29.62 / 13.44 / 9.80 yr (3.02×). At 800 km all three archetypes reach the integration cap. | every canonical orbit: P90 < P50 < P10 strictly AND P10/P90 > 2.0 | FAIL 3 of 4 orbits | PASS | JSON ↓ |
| M4 | Scenario-archetype invariance: P50 replayed from each complete solar cycle inside the P50 window (WDC-SILSO solar-cycle minima, month resolution). At 600 km the cycles give 15.48, 13.30, 15.93 and 21.44 yr. | rel_std <= 0.15 on every orbit whose four replays all end in reentry within 100 years | FAIL rel-std 0.181 at 600 km; 0 of 3 orbits | PENDING | JSON ↓ |
| M5 | /v1/lifecycle/mtbof endpoint live; auth gates; a Gone response for the legacy disposal route. Probed on the in-process API: neither route is registered. | exact response shape; 401 without token, 400 on bad input, 410 on the legacy route | NOT_IMPLEMENTED route answers 404 | PENDING | JSON ↓ |
| M6 | Envelope containment against documented historical reentries, each replayed forward from its end of mission: Skylab 5.42 yr in [2.61, 7.51]; UARS 5.78 in [5.57, 16.66]; ROSAT 12.69 in [8.01, 24.37]. The cohort's Skylab and UARS A/m values changed in the code after v0.4.1 (see What changed). | PASS if every reentry lies inside [P90, P10]; PASS_PARTIAL if some do; FAIL if none | PASS 3 inside, 0 outside | PASS_PARTIAL | JSON ↓ |
T-class — Cauchy/SO(3) tumbling-state observable T1 – T3
Tumbling state is treated as a topological observable on SO(3), with measurable secular drag enhancement via Cauchy's 1841 surface-area theorem. The methodology computes per-object α = ⟨A⟩isotropic / ⟨A⟩majoraxis and propagates a multiplicative correction to Lsigned for objects in the chaotic-tumble sector. The alternative dynamical-rectification mechanism was tested and not supported in v0.4.1 (null-test JSON); that test is not re-executed here, because its generator is not in the repository.
| ID | Claim | Tolerance | Result | v0.4.1 | JSON |
|---|---|---|---|---|---|
| T1 | Cauchy 1841: ⟨A⟩isotropic = S/4 for a convex body tumbling ergodically over SO(3). Checked against the closed-form surface and an independent quadrature over directions; Firefly Alpha 2nd stage: S = 39.02 m², Aiso = 9.75 m². | module A_iso vs S/4 rel-err <= 1e-12; quadrature vs S/4 rel-err <= 1e-9 | PASS quadrature rel-err ≤ 2.0e-11 | PASS | JSON ↓ |
| T2 | Major-axis-spin orbit-averaged baseline ⟨A⟩majoraxis = (2/π)·(Aendon + Abroadside), checked by independent quadrature on every cohort object | rel-err <= 1e-9 on every cohort object | PASS max rel-err 2.1e-11 | PASS | JSON ↓ |
| T3 | Canonical-cohort Tstate with leverage weighting: KH-8 (Gambit-3) reconnaissance bus +20.1%; Firefly Alpha 2nd stage +14.8%; Delta II 2nd stage (AJ10-118K) +12.7%; Shavit 3rd stage (RSA-3) +8.6%; Iridium-1 (decommissioned, LM-700 bus) -1.0% (panel-broadside-dominant geometry: the correction is not universally drag-enhancing). Every L_signed the cohort carries equals the pinned map's. | every L_signed the module carries equals the pinned leverage map's value for that cosparId; aggregate recomputed independently within 1e-12 | PASS α 0.99–1.20; weighted +15.84% | PASS | JSON ↓ |
Operational invariants Op1 – Op2
| ID | Claim | Tolerance | Result | v0.4.1 | JSON |
|---|---|---|---|---|---|
| Op1 | Snapshot pointer atomicity (decision #15): the API reads either the fully-written or the previous snapshot, never partial state. Contracts by kill injection. | all 5 contracts pass exactly (boolean assertions) | PASS 5/5 contracts | PASS | JSON ↓ |
| Op2 | Lookup contract: every lookup the slot route reads is written by the daily compile. safe_harbors_by_shell is read but never written: a snapshot from the daily compile alone serves the report without its safe-harbor layer. | set read is contained in set written (exact) | FAIL | new | JSON ↓ |
Engagement preview — methodology on public-catalog inputs
The artifacts below are demo data derived from the same substrate algebra and the same public DISCOSweb catalog used for the audited claims above, one per offering: Slot Assessment, MTBOF, LARS and the Substrate-Coupling Pathway. Engagements are direct. tony@fancyland.net.
| ID | Description | Notes | Result | JSON |
|---|---|---|---|---|
| SCP | Substrate-Coupling Pathway: the largest LEO constellation in the public catalog (11,422 active payloads) against the top-100 highest-|L_signed| objects that are NOT active payloads (rocket bodies, debris, fragments, decommissioned payloads), excluding the operator's own fleet. Primary sink: Firefly Alpha second stage 2023-202B (Rocket Body, 909 kg, 369 km, retrograde 140°) couples to 9,556 fleet members. Reproduces the v0.4.1 artifact: True. | exact (closed-form algebraic against the pinned leverage map SHA-256); the headline figures must equal the v0.4.1 artifact's, computed from the same map | DEMO_DATA 1,580 fleet members coupling above 5000 | JSON ↓ |
| INT | Closed-form Δ⟨Pσ⟩LEO under asset-set interventions. Removing the 80-object operator-fleet subset: Δ = -1.86e-06. Removing the 100 highest-|Lsigned| objects (1.502% of mass, 42.165% of |L|): Δ = -39.45%. Baseline agrees with C7a, equal when both are rounded to 6 decimal places (the precision the charter shows): True. Algebraic, not Monte Carlo. | exact (closed-form algebraic; reproducible to floating-point precision) | DEMO_DATA reproduces v0.4.1 PASS | JSON ↓ top-K ↓ |
| MTBOF | MTBOF P10 / P50 / P90 envelope replay across the canonical disposal orbits (600, 450, 800 and 410 km), with the M6 containment receipts above. | every canonical orbit: P90 < P50 < P10 strictly AND P10/P90 > 2.0 | DEMO_DATA 3.02× at 600 km | JSON ↓ |
| SWADI | Substrate-weighted in-situ atmospheric density per altitude shell: 9 populated shells across 450–1500 km; F factors 0.88–3.61. Bundled from the 2026-04-28 pipeline run; not recomputed. | coverage gaps surfaced, never extrapolated | DEMO_DATA | JSON ↓ |
Not re-verified in this edition
- The catalog itself. The DISCOSweb pull needs an ESA account token, which this run did not use. A fresh catalog would give a new leverage map and move every C-, S- and demo figure; C7b (drift of ⟨Pσ⟩LEO between catalogs) waits for it.
- TLE-based claims. A1, A2, A3, A5, A6 and A7, the real-TLE half of A13, and fresh A15, A16, A17 and SWADI data need the space-track.org TLE history, which needs an account login this run did not use.
- The deployed endpoint. S4, S7 and M5 ran on the API in-process; the deployed service and its production data were not called.
- Not built yet. M5's route is not registered in the API, and D3's lookup carries no cohort.
- No generator. The T-class dynamical-rectification null test has no generator in the repository and is not re-executed.
Reproducing this edition
Pinned inputs: leverage map leverage_map_v2_full_results.csv SHA-256 47f32fd6942384e282c2691bb4f8a0142be07f4843dfb10fc27dfcd7027cdd40; GFZ archive noaa_kp_ap_historical.csv SHA-256 6b811c8a0395abe6d97e613504414eaa9bb7c3702ae841f78c42690827474812, parsed from the GFZ source file SHA-256 551a3e483130409a42ab28d971425d68084da39da1217fd99b6ec6d32264d128; bundled SWADI run SHA-256 d485648389d42f259c864970b6931561510c6fbf56beb5a351d3c47556d7f5fd; bundled storm-response run SHA-256 2804047f84164075f91e201aed1922186625a608e093164b52665ba4d6b9b8f3. The edition is one command, python audit/run_audit.py; the file manifest with every JSON's SHA-256 is MANIFEST.json, and the row-by-row comparison with v0.4.1 is EDITION_DIFF.json.
Methodology source artifacts
Zenodo-deposited preprints and USPTO provisional patent applications. Each preprint is a permanent immutable artifact at its DOI. The Kessler compiler validates testable consequences of these published mathematics; the audit claims above cite specific results derivable from the listed sources.
| # | Preprint | Deposited | DOI |
|---|---|---|---|
| 1 | Prime Column Transition Matrix | 2026-03-17 | 10.5281/zenodo.19076680 |
| 2 | Separable Gyro-Bohm Scaling | 2026-03-19 | 10.5281/zenodo.19117880 |
| 3 | Spectral Isotropy | 2026-03-21 | 10.5281/zenodo.19156532 |
| 4 | Primorial Thermodynamics | 2026-03-23 | 10.5281/zenodo.19188924 |
| 5 | Universal Two-Prime Formula | 2026-03-24 | 10.5281/zenodo.19210625 |
| 6 | Active Transport on the Prime Gas / Arithmetic Qubit | 2026-03-27 | 10.5281/zenodo.19243258 |
| 7 | Arithmetic Black Hole | 2026-04-06 | 10.5281/zenodo.19442006 |
| 8 | The Unity Clock | 2026-04-08 | 10.5281/zenodo.19478727 |
| 9 | Patient Compass | 2026-04-13 | 10.5281/zenodo.19561701 |
| 10 | Character-Theoretic Effective Rank | 2026-04-24 | 10.5281/zenodo.19744573 |
| # | Provisional patent | Filed | Application |
|---|---|---|---|
| 1 | Arithmetic Qubit (state storage substrate) | 2026-04-07 | USPTO 64/031,440 |
| 2 | Holographic Eigen-Solver (spectral compute apparatus) | 2026-04-08 | USPTO 64/033,689 |
| 3 | Tensegrity Interferometer (stereoscopic measurement) | 2026-04-24 | USPTO 64/048,617 |