GuardianAI
Payment security for AI agents · live on Monad testnet

Spend caps say a payment is allowed.
GuardianAI checks it's right.

Before your agent pays, GuardianAI decodes what the call really does, checks who it pays and applies your rules for that agent. The agent's wallet won't move funds without GuardianAI's signature for that exact call, so even a fully fooled agent can't pay on its own.

12 contracts live on Monad testnet|271/271 Hardhat tests|2-of-3 Safe owns the contracts|Public drain-the-vault bounty

guardian://gate/live-demo
awaiting input

Architecture note: This sandbox runs the fast-path screen and de-obfuscation in your browser. The semantic classifier, output redaction and the on-chain PolicyGuard check run in the full gateway; try them in the Attack Lab.

Unlimited token approval Permit2 approval Role grant Zero-slippage swap Scam-listed recipient Lookalike address Forged approval Impostor agent Over-cap payment Stolen agent key Hidden instructions
0

funds an agent can move
without GuardianAI's signature

12

contracts live on Monad testnet
check them live ↗

271/271

Hardhat contract tests passing
test snapshot

2-of-3

Safe multisig owns the contracts
since 5 Nov 2026

DISCLOSURE: Compliance and audit artifacts are internally verified empirical test suites (3,211 prompts across 8 corpora, 1,490+ Python & 271 Hardhat tests, 192 Mera Passkey PRF assertions), not third-party statutory certifications.

Threat model

Four checks before GuardianAI signs a payment.

A payment is signed only if all four pass. The wallet checks that signature on-chain, so the agent can't skip GuardianAI, however it's been tricked.

[01]

What the call really does

The calldata is decoded, not trusted. Unlimited ERC-20 and Permit2 approvals, role grants, zero-slippage swaps and reserve manipulation are caught even when they target a contract you allow.

[02]

Who it pays

Recipients are checked against GuardianAI's threat feed before anything is signed. The guarded-wallet middleware also flags lookalikes of addresses you trust or have paid recently, a common way to poison an agent's address book.

[03]

Your rules for this agent

Per-agent rules in the relay, plus daily outflow caps and a contract and function allowlist enforced by the wallet itself on-chain.

[04]

Who is asking

Each agent carries a passkey-signed ID card and an ERC-8004 passport. Strict mode refuses any agent that isn't registered, so an impostor can't borrow a real agent's limits.

How it fits together

Checked off-chain. Enforced on-chain.

The relay decides whether a payment is right and signs it. The wallet refuses anything unsigned or over its limits, so a failure on one side doesn't open the other.

off-chain // the decision
  • Payment checks — decoded calldata, threat-listed recipients and per-agent rules, scored before anything is signed
  • Signer in Cloud KMS — the attestation key can live in Google Cloud KMS, so it never enters the relay process
  • Single-use approvals — each signature is bound to one call's value, calldata and nonce, and can't be replayed
  • Input firewall (extra layer) — screens prompts for injection before they reach the model; useful, but the payment check is the guarantee
For agent developers: Your agent keeps its own reasoning code. It asks GuardianAI to approve a payment, then submits it through the guarded wallet. The SDK and ElizaOS plugin handle both steps.
on-chain // the enforcement
  • GuardianAgentWallet — funds move only with the agent's key and GuardianAI's signature, under daily outflow caps and a contract and function allowlist (live on Monad testnet; operator is a policy-locked Privy server wallet)
  • PolicyGuard — verifies GuardianAI's signed approval and refuses revoked agents (live on Monad testnet)
  • Agent identity — a Mera passkey signs for each agent and an ERC-8004 passport records it; GuardianAI only approves payments for agents the passkey signed for. Try it with your passkey
In plain English: If the agent is fooled, GuardianAI won't sign. If GuardianAI were fooled, the wallet's caps and allowlist still hold. If the agent's key is stolen, nothing moves without the signature.
Building next · not live yet

Where GuardianAI is going.

Everything above runs today on Monad testnet. These are next, and we'll mark each one live only when you can try it.

[→]

Works with the wallet you already use

GuardianAI as a policy co-signer for Privy, Turnkey, Coinbase CDP and Safe, so you add the check without moving funds to a new wallet. Base support comes with it.

[→]

Checks the payment against the task

Each payment is simulated and compared with what the agent was asked to do ("buy API credits from X, up to $20 a day"), not only with static limits.

[→]

Asks you instead of only blocking

Payments outside policy go to the owner for a one-tap passkey approval, so a strict policy doesn't stall a legitimate agent.

[→]

Every agent protects the others

Scam recipients, malicious payment endpoints and drain patterns seen by one agent are refused for all of them.

Evidence

Numbers with receipts. Including the ugly ones.

Every figure on this site cites its dated repository artifact — and the proof page publishes full benchmark tables, not just highlights.

[A]

Benchmarks →

3,211 prompts across 8 public datasets, August 2026. AdvBench 99.0% strict, JBB PAIR+GCG 98.7% — and Do-Not-Answer's 57.1% published too, because selective reporting is how trust dies.

[B]

Audit status →

One external review of the off-chain gateway (March 2026, summary published, zero open critical/high). Internal off-chain, contract and architecture audits July–September 2026, including the Mera passkey enclave audit (166/166 assertions). The Monad contracts have not been externally audited yet.

Get started

Your agent can already pay.
Make sure it checks first.

Running agents that hold real funds? Tell us what nearly went wrong. We're choosing the next integrations based on it.