DYOR

Problem

Today's AI agents can search, trade, and automate workflows, but they often trust user input without verifying whether a Web3 project, token, contract, or repository is actually the intended entity.

That creates serious risks:

• interacting with fake tokens

• using phishing domains

• analysing the wrong repository

• executing trades on malicious contracts

• making financial decisions using incomplete information

Existing research tools generate reports.

They do not act as a verification layer for autonomous agents.

Solution

DYOR is a trust infrastructure layer for AI agents.

Before an agent performs a financial or blockchain action, it asks DYOR:

"What exactly is this?"

DYOR:

• resolves canonical identity

• validates the entity across multiple trusted providers

• detects ambiguous identities

• blocks unknown projects

• gathers live evidence

• evaluates risk

• returns structured machine-readable decisions

Instead of long AI-generated reports, agents receive deterministic outputs such as:

VERIFIED
RESOLVED
AMBIGUOUS
UNRESOLVED

PROCEED
REVIEW
BLOCK

Key Features

Canonical Entity Resolution

Maps domains, GitHub repositories, token symbols and contract addresses into a single canonical project identity.

Confidence Engine

Every resolution includes explainable confidence rather than blind assumptions.

Ambiguity Detection

Instead of guessing, DYOR blocks ambiguous inputs and requests clarification.

Live Multi-source Evidence

Collects evidence from trusted Web3 providers.

Structured Risk Engine

Returns structured data for AI agents rather than human-only reports.

Agent-native API

Designed for autonomous agents rather than traditional dashboards.

Why it matters

The next generation of AI agents will manage wallets, execute trades, deploy contracts and move capital.

DYOR ensures they verify before they act.

Week 1

Designed architecture for an AI-native Web3 verification engine.

Implemented canonical entity resolution.

Week 2

Built deterministic input classification.

Implemented confidence scoring.

Added canonical identity graph.

Week 3

Implemented ambiguity detection.

Blocked unresolved entities before provider execution.

Added explainable confidence.

Week 4

Integrated live provider evidence.

Built structured API responses.

Optimised repository and contract resolution.

Current

Completed a production-style verification pipeline capable of resolving known identities while safely blocking ambiguous or malicious inputs.