Introduction

Spellguard is the control plane for managing AI agents across your organization — multi-vendor, multi-department, unified governance. Get visibility into every agent, enforce policies consistently, and maintain audit trails across your entire agentic infrastructure.

For secure agent-to-agent communication, Spellguard provides two core protocols that you can use independently: cTLS and AMP.

cTLS

Cloud Transport Layer Security implements bidirectional attestation and secure channel establishment for agents regardless of the hosting provider.

AMP

Admissible Messaging Policy provides tamper-evident audit logs and secure message archiving between two parties without requiring trust of either side's account.

Key Properties

• cTLS
• AMP

• Bidirectional verification before communication is exchanged
• Forward secrecy through ephemeral RAM-only keys
• Cryptographic proof of verifier privacy and neutrality

• Commitments prove existence without revealing content
• Pluggable backends (S3, Rekor, others)
• Bilateral and unilateral attestation modes