Skip to content

VVS-1 Overview

VVS-1 (Venmail Verification Standard, version 1) adds verifiable, agent-scoped identity and message integrity to email. It uses Ed25519 signatures, standard SMTP headers, and existing DNS/HTTPS infrastructure.

VVS-1 does not modify SMTP. It layers trust on top of delivery — receivers that don’t support VVS-1 deliver messages normally.

Email has no native way to verify that a message from [email protected] was actually sent by the billing system (and not spoofed, injected, or tampered with). DKIM proves the domain sent it. VVS-1 proves which agent sent it and that the body hasn’t been modified.

This matters for:

  • AI agents sending emails on behalf of users or systems
  • Automated workflows where message authenticity is critical
  • Financial communications that must be tamper-evident
  • Cross-organization agent-to-agent communication
  1. Each agent gets an Ed25519 keypair
  2. The public key is published at /.well-known/venmail-agent/{name} or via DNS
  3. When sending, the agent signs the message body + headers
  4. VVS headers are added to the email before DKIM signing
  1. Extract X-Venmail-* headers from the incoming message
  2. Resolve the sender’s public key via well-known, DNS, or embedded header
  3. Verify the Ed25519 signature against the message content
  4. Assign a trust level: VERIFIED, PARTIAL, FAILED, or UNKNOWN
LevelMeaning
VERIFIEDSignature valid, key resolved via .well-known or DNS
PARTIALSignature valid, key from embedded header only
FAILEDHeaders present but verification failed
UNKNOWNNo VVS headers — normal email
  • Messages without VVS headers are delivered normally
  • VVS verification never blocks SMTP delivery
  • Trust levels are annotations, not gatekeepers