PROTOCOL_EVIDENCE // V1.0_GENESIS

Show Me

Verifiable proof that MyShape Protocol distinguishes human presence from synthetic simulation. Every claim on this page is reproducible.

PES Benchmark Results

Test Case	PES	Verdict
Genuine Human	0.9817	PASS ✓
AI Forgery (GPT-5)	0.5857	FAIL ✗
Impostor	0.0000	FAIL ✗

Human—AI Gap0.3960

Source: myshape-demo CLI · Rust core engine · 25/25 tests pass

Identity Vector Sample

KinematicsBone length ratios

0.42

AccelerationHurst exponent

0.88

JerkMAD + lag-1

0.15

Jerk Spectrum1/f scaling

0.71

PES ScoreComposite

0.94

ZK HashSHA-256

0x7f3a...b2e1

Attack Surface Summary

A — Generative AIFrequency entropy collapse

~0%

B — ReplayUniform inter-frame timing

Near zero

C — Sensor SpoofMulti-dim corroboration

Very low

Live Verification Engine

LIVE_VERIFICATION // v0.1

Presence Verification Dashboard

Runtime engine output from myshape-demo CLI. Motion Signature Engine (128-dim, 4 feature groups) + Multi-Factor Scorer. All scores computed by the Rust core engine compiled to native binary.

HUMAN—AI DIVERGENCE

0.0000

Presence Gap

The irreducible distance between genuine biological motion and the best AI-generated forgery. This gap exists because AI models cannot replicate the physics of the human nervous system.

GENUINE HUMAN

0.0000

PRESENCE CONFIRMED

Real-time biological motion · 8–12 Hz tremor · 1/f jerk spectrum

Motion

0.970

Device

1.000

Context

1.000

Threshold0.75

AI FORGERY

0.0000

PRESENCE DENIED

GPT-5 / DeepSeek synthetic motion · over-smoothed dynamics

Motion

0.310

Device

1.000

Context

1.000

Threshold0.75

IMPOSTOR

0.0000

HARD GATE · DEVICE MISMATCH

Different human entity · different device · kinematic mismatch

Motion

0.955

Device

0.000

Context

0.000

Threshold0.75

AI REJECTION ROOT CAUSE ANALYSIS

The AI forgery failed verification because the Motion Signature Engine detected structural deficiencies across four independent dimensions. These are not temporary AI limitations — they are consequences of how neural network architectures model motion.

Enginev0.1.0 · 25/25 tests pass

FeaturesK · A · J · J_spec (4/7 active)

FactorsM(0.60) · D(0.25) · C(0.15)

Risk APILOW · MEDIUM · HIGH

All evidence is reproducible. Run the benchmarks yourself.

Try Live Demo →Full Threat Model →Whitepaper →

SYNCHRONIZING

PROTOCOL_EVIDENCE // V1.0_GENESIS

Show Me

Verifiable proof that MyShape Protocol distinguishes human presence from synthetic simulation. Every claim on this page is reproducible.

PES Benchmark Results

Test Case	PES	Verdict
Genuine Human	0.9817	PASS ✓
AI Forgery (GPT-5)	0.5857	FAIL ✗
Impostor	0.0000	FAIL ✗

Human—AI Gap0.3960

Source: myshape-demo CLI · Rust core engine · 25/25 tests pass

Identity Vector Sample

KinematicsBone length ratios

0.42

AccelerationHurst exponent

0.88

JerkMAD + lag-1

0.15

Jerk Spectrum1/f scaling

0.71

PES ScoreComposite

0.94

ZK HashSHA-256

0x7f3a...b2e1

Attack Surface Summary

A — Generative AIFrequency entropy collapse

~0%

B — ReplayUniform inter-frame timing

Near zero

C — Sensor SpoofMulti-dim corroboration

Very low

Live Verification Engine

LIVE_VERIFICATION // v0.1

Presence Verification Dashboard

Runtime engine output from myshape-demo CLI. Motion Signature Engine (128-dim, 4 feature groups) + Multi-Factor Scorer. All scores computed by the Rust core engine compiled to native binary.

HUMAN—AI DIVERGENCE

0.0000

Presence Gap

The irreducible distance between genuine biological motion and the best AI-generated forgery. This gap exists because AI models cannot replicate the physics of the human nervous system.

GENUINE HUMAN

0.0000

PRESENCE CONFIRMED

Real-time biological motion · 8–12 Hz tremor · 1/f jerk spectrum

Motion

0.970

Device

1.000

Context

1.000

Threshold0.75

AI FORGERY

0.0000

PRESENCE DENIED

GPT-5 / DeepSeek synthetic motion · over-smoothed dynamics

Motion

0.310

Device

1.000

Context

1.000

Threshold0.75

IMPOSTOR

0.0000

HARD GATE · DEVICE MISMATCH

Different human entity · different device · kinematic mismatch

Motion

0.955

Device

0.000

Context

0.000

Threshold0.75

AI REJECTION ROOT CAUSE ANALYSIS

Enginev0.1.0 · 25/25 tests pass

FeaturesK · A · J · J_spec (4/7 active)

FactorsM(0.60) · D(0.25) · C(0.15)

Risk APILOW · MEDIUM · HIGH

All evidence is reproducible. Run the benchmarks yourself.

Try Live Demo →Full Threat Model →Whitepaper →