top of page

Give Your Agent a Pulse

Sep 29
7 min read

Today we are launching Yanez Pulse: biometric authorization for people and the AI agents that act on their behalf. Pulse ships with Gate, a feature set that puts an air gap between your agent and your wallet.



Would you give an AI agent your wallet? Your credit card?


That question sits underneath one of the largest shifts in commerce in decades. McKinsey estimates that AI agents could mediate $3 trillion to $5 trillion of global consumer commerce by 2030. Every dollar of that depends on a person giving an agent the authority to act on their behalf.


Today, people hesitate to do that. Capgemini surveyed 1,500 executives at companies with more than $1 billion in revenue and found that trust in fully autonomous AI agents fell from 43% to 27% in a single year. On the consumer side, Checkout.com's 2026 research found that one in four shoppers say they will never delegate a purchase to AI, and 27% trust no organization to operate an AI shopping agent.


That hesitation is rational. Agents can go rogue. The technology to keep them in check exists, and today we are shipping it. It's called Pulse.


What "rogue" looks like


The organizations closest to the problem are candid about it. OpenAI wrote that prompt injection, "much like scams and social engineering on the web, is unlikely to ever be fully 'solved.'" The UK's National Cyber Security Centre says it "may never be totally mitigated." Gartner predicts that by 2028, 25% of enterprise breaches will be traced back to AI agent abuse. The OWASP Top 10 for Agentic Applications lists Identity & Privilege Abuse and Rogue Agents among the defining risks of the category.


If agents will sometimes be tricked, the useful question is what a tricked agent can actually do. Pulse was built to answer it.


Two root problems

I have spent more than 30 years in fraud and financial crime prevention. Through that lens, two structural issues show up again and again.


1. Human authorization is proxied by something you own. API keys, OAuth tokens, session cookies and passwords all stand in for the person. Whoever or whatever holds the token is, as far as the system can tell, you. These tokens leak at scale. GitGuardian counted 28.65 million new hardcoded secrets on public GitHub in 2025, with AI-service secrets up 81% year over year and more than 24,000 secrets exposed in MCP configuration files alone. 64% of valid secrets found in 2022 were still exploitable in January 2026. An agent carrying a proxy credential carries your authority wherever it goes.


2. The agent holds the payment details. Most agent payment setups today give the agent a wallet, a private key or a card credential so it can transact. Simon Willison's "lethal trifecta" describes the result: an agent with access to private data, exposure to untrusted content and the ability to communicate externally can be steered into moving what it holds. In the Lobstar Wilde and Grok cases, the agent held the funds it lost.


Pulse addresses both problems together.


How Pulse works

Pulse is biometric authorization for humans and their transactions, with a focus on the transactions agents make on our behalf. It does two things.


1. Authorization that comes from you



The flow:

  1. Your agent requests a sensitive action.

  2. Pulse pings your phone.

  3. You Pulse (verify with your face) in the Yanez YID app.

  4. A signed approval goes back, and the agent executes.


You decide when the agent requires authorization. Routine actions run on their own, and high-value or sensitive actions wait for you. A trading agent, for example, can track a portfolio and propose trades on its own, while any trade above your threshold waits until your face approves it on your phone.


The approval comes from your live biometric presence at the moment of the request. There is no password, token or key that stands in for you, so a leaked credential gives an attacker nothing to approve with.


2. Gate: an air gap between your agent and your wallet


Pulse makes sure only you can say yes. With Gate, Pulse also takes the payment details away from the agent entirely.



Gate is the Pulse feature set for payments. At its core is a smart contract you configure, which ties your payment wallet and your policy, including spending limits, to your Pulse authorization. The funds sit behind the contract, and the agent never holds a wallet, a key or a card number.


When you approve an action, your face produces a signature at that moment, bound to that exact action and policy. The contract verifies that proof on-chain before anything moves. Every other path is refused.


Here are three ways to attack a Pulse-protected wallet, and what happens:


  • Trick the agent into bypassing Pulse. The funds sit behind the contract, so the agent has nothing to release.

  • Trick the agent into firing requests over and over. Every request asks the owner for a live biometric approval. So you are unlikely to approve rogue transactions over and over.

  • Broadcast a transaction directly, outside the agent. Anyone can relay a transaction. The contract unlocks only for a signature derived from the owner's biometrics, bound to that exact action and policy.



An agent can't spend what it doesn't have access to. With Pulse and Gate in place, the transfers in the Grok and Lobstar Wilde incidents would each have required a live biometric approval from the owner before any funds moved.


In Checkout.com's research, spending caps (30%) and instant revocation (29%) were among the top safeguards shoppers want before delegating purchases to an agent. Pulse puts both in the owner's hands, enforced by code.


Built on Yanez YID

Pulse runs on Yanez YID, our Proof of Humanhood and Proof of Uniqueness (PoH/PoU) ecosystem. Yanez PoH/PoU is built on three principles:


  • Privacy-preserving. Biometrics never leave your device and are never stored.

  • Hardware agnostic. YID works on the phones and hardware people already own. Our rollouts with dFusion and Syklo already run on hundreds of device types worldwide. Solutions that depend on dedicated scanners or specific hardware can only reach the people who have access to that hardware.

  • Decentralized. A transparent ecosystem that anyone can build on.


Proof of Uniqueness matters for agents as well. It lets a service confirm that one real human stands behind an approval, and that it is the same human every time.



Who we built this for

Teams building enterprise agentic infrastructure. If your organization is deploying agents that touch payments, treasury, procurement or customer accounts, Pulse gives you a human authorization step that no credential can impersonate, and keeps funds outside the agent's reach with policy enforced on-chain.


Developers. If you are building agents or agent platforms, Pulse lets you offer your users biometric approval and gated spending without building identity infrastructure or handling biometric data yourself.


Vibe coders. If you are building your own agents and wondering what happens when one of them misreads a prompt with your money attached, connect Pulse to your agent, turn on Gate, and set your limits.


Availability and pricing

  • Pulse is available today. Download Yanez YID from the App Store or Google Play and connect Pulse to your agent. It is free until the end of October.

  • Gate, Pulse's payment protection, is available for configuration now, priced at 1% of the transaction it protects.

  • Get started: see the agents and apps already running on Pulse and YID in the Pulse showcase, part of the Yanez developer docs.


Agents are going to handle a large share of how we buy, trade and pay. Some of them will be tricked, and some will go rogue. We built Pulse so that when it happens, the agent has no wallet to drain and no credential that can approve on your behalf. Only you can say yes.


Give your agent a Pulse.


Jose Caldera, Founder & CEO, Yanez


Sources

  1. McKinsey, "The automation curve in agentic commerce" (Jan 2026), citing "The agentic commerce opportunity" (Oct 2025). https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-automation-curve-in-agentic-commerce

  2. Capgemini Research Institute, "Rise of agentic AI" press release (July 2025). https://www.capgemini.com/news/press-releases/trust-and-human-ai-collaboration-set-to-define-the-next-era-of-agentic-ai-unlocking-450-billion-opportunity-by-2028/

  3. Checkout.com, "Agentic Commerce 2026" (June 2026). https://www.checkout.com/newsroom/consumer-demand-for-ai-shopping-is-forming-fast-but-trust-for-agentic-commerce-is-still-catching-up

  4. SailPoint / Dimensional Research, "AI agents: The new attack surface" (May 2025). https://www.sailpoint.com/press-releases/sailpoint-ai-agent-adoption-report

  5. Fortune, Replit production database deletion (July 2025). https://fortune.com/2025/07/23/ai-coding-tool-replit-wiped-database-called-it-a-catastrophic-failure

  6. Cointelegraph, Lobstar Wilde transfer (Feb 2026). https://cointelegraph.com/news/openai-employee-s-ai-agent-accidentally-sent-442k-to-beggar

  7. OECD.AI Incident Database, Grok/Bankrbot prompt injection (May 2026). https://oecd.ai/en/incidents/2026-05-04-4a73

  8. Giskard, "How Grok got prompt-injected" (May 2026). https://www.giskard.ai/knowledge/how-grok-got-prompt-injected-an-x-user-drained-150-000-from-an-ai-wallet

  9. Gambit Security, autonomous AI agents against online retailers (Sept 2026). https://gambit.security/blog-posts/autonomous-ai-agents-online-retailers-25-a-company

  10. TechCrunch, OpenAI and UK NCSC on prompt injection (Dec 2025). https://techcrunch.com/2025/12/22/openai-says-ai-browsers-may-always-be-vulnerable-to-prompt-injection-attacks/

  11. Gartner, Top Predictions for IT Organizations and Users in 2025 and Beyond (Oct 2024). https://www.gartner.com/en/newsroom/press-releases/2024-10-22-gartner-unveils-top-predictions-for-it-organizations-and-users-in-2025-and-beyond

  12. OWASP GenAI Security Project, Top 10 for Agentic Applications (Dec 2025). https://genai.owasp.org/2025/12/09/owasp-top-10-for-agentic-applications-the-benchmark-for-agentic-security-in-the-age-of-autonomous-ai/

  13. GitGuardian, "The State of Secrets Sprawl 2026." https://blog.gitguardian.com/the-state-of-secrets-sprawl-2026/

  14. Simon Willison, "The lethal trifecta for AI agents" (June 2025). https://simonwillison.net/2025/Jun/16/the-lethal-trifecta/

  15. CBS News, OpenAI agents and U.S. government websites (Sept 2026). https://www.cbsnews.com/news/openai-ai-agent-bot-rogue-hack-government-website/


bottom of page