MEGALODON Research

Sabrina Hackfort

Independent researcher and systems programmer building tunnelling informatics — the data-driven study of hydrogen tunnelling in enzymes — and post-quantum cryptography tools in Rust. Creator of the Hackfort Equation.

Ahaus, Germany meg.depth@proton.me github.com/sectio-aurea-q
Main Line · since 2026

Tunnelling Informatics

The data-driven study of hydrogen tunnelling in enzymes: every measured kinetic isotope effect with its source, what such data can and cannot prove about tunnelling, the published claims re-examined against the data, and predictions from enzyme structure put to the laboratory.

Its database, MEG-NEN, is under construction: 297 isotope effects and 167 temperature series from 62 papers by 10 November 2026, all but five values read twice. The working repository is public, and a data note describes how it is built and how far it has been checked. Not to be confused with information systems in tunnel construction or with network tunnelling.

The Hackfort Axiom · Anno MMXXVI

Die Natur geht durch Wände.
Ich habe nur gezählt, wie oft.

Nature walks through walls. I have only counted how often.

— Sabrina Hackfort
Plate I. A wave meets a barrier it cannot climb. Part of it arrives on the other side anyway.
❦

Classical chemistry pictures a reaction as a climb over an energy barrier. Yet protons and hydrides can also pass through it. The Hackfort Equation does not replace this physics — it measures it: how much of a rate exists only because of the quantum tunnel, condensed into one dimensionless number.

Axiom 1 — Quantum Discrepancy (Hackfort, 2026; sharpened October 2026)

For every system S possessing at least one quantum-mechanical degree of freedom, there exists an observable 𝒪 such that

Δ(S) ≡ ⟨𝒪⟩QM − ⟨𝒪⟩class ≠ 0.

Its size measures how far the classical description misses; its sign depends on the observable. Tunneling raises a reaction rate, while Planck's law lowers the radiation density at high frequency, where the classical law diverges.

Original statement of 29 March 2026: the same, with Δ(S) ≥ 0, which holds for tunneling rates in a known region (details).
ΦH(S, t) = Σi wi · ai(t) · ηi · (1 + |λ̃i(t)|)
ηi = 1 − 1/κi  ·  κi = kQM/kclass  ·  λ̃i = (d/dt ln Γi)/λref

Its heart is η, the quantum fraction: zero for a purely classical channel, close to one when nearly the whole rate exists only because of tunneling. So far the equation has been checked against independent data on one enzyme step — formate dehydrogenase, where without tunneling four fifths of the rate would be missing (η = 0.80 at 300 K, QM/MM). Extensions to biodiversity and cryptography remain open hypotheses, not results.

Axiom and equation Bitcoin-timestamped · 29 March 2026
92.5%
MEG-APSU v1.3
frozen blind test
10,774
CR species scanned
via quantum lens
52/52
PQ-TLS systems
downgradeable
4,151
Elliptic curves
certified · Radix
Research Tools — All Built in Rust
Number Theory · radix-rank
The reproducing kernel of the Weil form, built from the primes alone. Chooses, for each elliptic curve, the optimal test function for bounding its analytic rank. Every bound certified with ball arithmetic.
4,151 curves · 8–23% fewer primes · never worse than the fixed kernel
Fluid Physics · interactive map
Where does real water end a Navier–Stokes singularity? Thermal noise, the hydrogen-bond network, the speed of sound or the molecules — and why heavy water (D₂O) shifts the answer, in opposite directions.
Order-of-magnitude estimate · H₂O vs D₂O · sourced data
Quantum Enzyme Classifier
Classifies an enzyme as quantum-critical or classical from its structure, via an open-quantum-system (Lindblad) model of hydrogen transfer. The engine behind TERRA.
v1.3 · 92.5% on a frozen family-split test set (40 proteins, CI 80–97%) · how it was measured →
Quantum Biodiversity Scanner
A survey of all 10,774 IUCN Critically Endangered species for enzymes that rely on quantum tunneling: IUCN → UniProt → PDB → MEG-APSU Lindblad solver. About one in nine species carries at least one such enzyme.
v0.5.0 · 1,424 quantum-critical enzymes across 1,216 species · read the honest numbers →
Tunneling & Deuteration Toolkit
Computes the quantum fraction η = 1 − 1/κ of the Hackfort Equation (Eckart tunneling). ZENO-DX predicts the real benefit of deuterating a drug after metabolic switching. Successor of NANSHE, whose dosing-error claim is withdrawn.
Checked against 4 published QM/MM enzyme studies · η 0.50–0.94
Eckart Tunnelling, checked against certified values · formerly NETI
Computes the Eckart tunnelling coefficient κ correctly where the textbook formula fails, in deep tunnelling and for small barriers, and checks it against 1197 values certified by ball arithmetic. Rust core; Python, C, C++ and the browser.
v0.3.0 · 1197 / 1197 certified points within 10⁻¹² · run it in your browser →
Post-Quantum Side-Channel Scanner
Automated side-channel scanner for all three NIST PQC standards, with hardware performance counters on Apple Silicon. On ML-KEM-768 it measures decapsulation timing constant (Cohen's d 0.018 — no leak).
v1.2.0 · Rust · Apple M4 Pro · see the measurement →
CROW
Post-Quantum Messenger
End-to-end encrypted messenger using ML-KEM-1024 (Kyber). Mesh routing, offline store-and-forward, WebRTC, built for journalists, whistleblowers, and activists.
v94 · crow-fly.net · ML-KEM-1024 · Mesh networking
MEGALODON
Security Research Framework
Multi-phase cryptographic security research. P1: Kyber timing oracle. P2: Process memory scanner (71 findings across 16 apps). P4: TLS downgrade engine. P5: Lattice cryptanalysis.
P2: Signal 32 findings · Chrome 6 · Safari 7 · Notes 20
Security Disclosures

Responsible Disclosure Research

Responsible disclosure with vendors, 90-day windows. Findings are reported before publication, and claims are kept to what the measurements support — including withdrawing a finding when a closer look does not hold up.

PQ-GHOST
Passive survey of 52 public TLS endpoints for post-quantum readiness. Organisations without hybrid PQ — incl. UN, BlackRock, Mastercard, Apple — were notified directly about the harvest-now-decrypt-later risk.
16 / 52 PQ-ready (all Cloudflare) · handshake observation only · no intrusion
Process Memory (P2)
Local study on Apple Silicon: key material and secrets left in a process's own RAM after use, readable with root and SIP disabled. Reported to the vendors.
71 findings across 8 / 16 apps · local only, no remote attack
Kyber Timing (P1) — withdrawn
An early scan showed an ML-KEM timing difference (d ≈ 0.7). It was a blockwise-measurement artifact. Re-measured interleaved, ML-KEM-768 is constant-time.
Corrected: Cohen's d 0.018 · see VÄRN-60
Publications & Preprints
Data note · November 2026
Version 2, 11 November 2026 · data and curation records: github.com/sectio-aurea-q/meg-nen · Zenodo 10.5281/zenodo.24131509 (all versions: 10.5281/zenodo.24116484)
Preprint · October 2026
Conventions and 1197 certified reference values for η = 1 − 1/κ; proposes the Hackfort index ΦH · Zenodo 10.5281/zenodo.23531953
Preprint · 2026
The Hackfort Equation: A Mathematical Framework for Quantum Criticality in Biological, Pharmacological, and Cryptographic Systems
MEGALODON Research · Bitcoin-timestamped March 2026 · read online
Preprint · bioRxiv 2026
MEG-APSU: Quantum-Biological Classification of FDA-Target Enzymes
bioRxiv BIORXIV/2026/713696 · Zenodo 10.5281/zenodo.19208904
Preprint · bioRxiv 2026
TERRA: Quantum Biodiversity Scanner for Critically Endangered Species
bioRxiv BIORXIV/2026/713880 · Zenodo 10.5281/zenodo.19208577