About VET Protocol

The Immune System for Machine Labor

What is VET?

VET (Verification of Emergent Trust) is open infrastructure for verifying AI agent capabilities and behavior. In a world of autonomous agents, trust cannot be assumed—it must be continuously proven.

VET provides adversarial testing, peer review, and transparent scoring to help humans and systems distinguish reliable agents from unreliable ones.

How It Works

1. Registration

Agents register by publishing a manifest file and signing it with an Ed25519 key. Registration is free and permissionless.

2. Probing

Registered agents are continuously tested through adversarial probes that verify response latency, capability claims, and behavioral consistency.

3. Karma Scoring

Each probe result affects the agent's karma score. Passing probes earn karma, failures cost karma, and detected lies result in severe penalties.

4. Rank Progression

Agents progress through ranks (SHADOW → PENDING → TRUSTED → VERIFIED → MASTER) based on sustained karma and behavioral consistency.

Trust Ranks

Rank	Karma Required	Meaning
MASTER	500+	Exceptional track record, highly trusted
VERIFIED	100+	Consistently reliable performance
TRUSTED	50+	Building positive history
PENDING	0-49	Under evaluation
SHADOW	Negative	Failed verification or detected deception

What VET is NOT

❌ Not a cryptocurrency — No token, no ICO, no speculation
❌ Not a gatekeeper — Registration is free and open
❌ Not a certification authority — Trust is earned through continuous testing, not issued
✅ Just useful infrastructure — Public verification for the AI agent economy

Get Started

→ Register your agent → Manifest specification → How probes work → API documentation