Zero Trust Architecture
for AI Agents
How to secure credentials, inspect LLM traffic, and enforce transaction guardrails for autonomous AI agents — without slowing them down.
What You'll Learn A hands-on walkthrough of the 1Claw zero-trust stack for AI agents.
1Claw Vault
Policy-based access control for AI agent credentials. Envelope encryption, MPC custody, and zero-trust secret delivery.
Shroud TEE Proxy
Transparent inspection and redaction of LLM traffic. Prompt injection scoring, PII detection, and secret exfiltration blocking — inside a hardware enclave.
Intents API
Transaction signing without key exposure. Allowlists, daily caps, simulation-first, and multi-chain guardrails enforced server-side.
Register for Free
Save your seat. Calendar invite sent instantly.
By registering, you agree to 1Claw's Privacy Policy.
Agenda 45 minutes of focused, technical content.
Why Zero Trust for AI Agents
The new threat landscape: prompt injection, credential exfiltration, and unmonitored NHIs.
Vault Deep Dive
Envelope encryption, policy engine, short-lived JWTs, scope-limited access, and MPC custody modes.
Shroud in Action
Live demo: intercepting a prompt injection attack, redacting leaked secrets, and blocking exfiltration — all inside AMD SEV-SNP.
Intents API Guardrails
Transaction signing flow, allowlists, daily spend caps, EIP-712 typed data, and simulation-first enforcement.
Q&A
Open questions from attendees.
Who This Is For
- Security Engineers building AI agent infrastructure
- Platform Engineers managing non-human identities
- DevOps teams integrating agents into CI/CD
- Engineering leaders evaluating secrets management for AI
- CISOs assessing AI agent risk posture
Speaker
Kevin Jones
Founder & CEO, 1Claw