Skip to content

README.md

Latest commit

 

History

History
34 lines (28 loc) · 830 Bytes

README.md

File metadata and controls

34 lines (28 loc) · 830 Bytes
Raw

AgentGuard

A GenLayer policy gateway that verifies AI agent actions against user rules and live evidence before execution.

AgentGuard adds a verifiable safety layer between AI agents and sensitive actions.

Users provide:

  • A policy describing the rules the agent must follow
  • A proposed agent action
  • An HTTPS evidence source

The AgentGuard Intelligent Contract retrieves the external evidence, evaluates the action against the policy, reaches a GenLayer consensus decision, and records the verification result on-chain.

How It Works

User
  ↓
AgentGuard Web App
  ↓
Policy + Agent Action + Evidence URL
  ↓
GenLayer Intelligent Contract
  ↓
Contract-side HTTPS Evidence Retrieval
  ↓
Policy + Evidence Evaluation
  ↓
GenLayer Consensus
  ↓
ALLOW / BLOCK
  ↓
On-chain Verification Record
E0F