Give an AI agent limited spending rights on your wallet
The safest way to let an AI agent spend from your wallet is a permission that a contract enforces: one action, one token, a limit per use, a lifetime limit and an expiry, which you can revoke at any time. In Signo this is an Execute permission, enforced by the Signo Shield contract on X Layer and Arbitrum; the agent never gets your key and never holds your tokens.
Last verified: 16 October 2026. Execute is in early access.
Signo gives you AI agents that watch your DeFi positions around the clock. Each agent follows a position or a market condition you choose, alerts you on Telegram when it matters, and prepares the next step for you to review and sign in your own wallet. Signo never holds your funds. In early access, agents can also act on their own within limits you sign onchain and can revoke at any time. This page is about that last part.
What you sign
| Part | What it means |
|---|---|
| Action | The one thing the agent may do: swap, send, deposit into or withdraw from a vault, claim rewards, or on Aave supply, repay or repay from collateral. |
| Token | The one token the contract may take from your wallet. |
| Per-firing limit | The most one firing may spend, before the fee. If you leave it empty, there is no per-firing limit. |
| Lifetime limit | The total across all firings, before fees. It does not reset. If you leave it empty, there is no lifetime limit. |
| Expiry | After this time, the permission cannot fire. If you leave it empty, the expiry is five years. |
| Venues | The exact contracts the action may call. |
| Trigger | The part of your condition the chain can read: a health factor, a token balance or a price. |
| Required result | What must be true after the action, for example "your debt fell". |
What the contract checks at every firing
- Only Signo's agent key fires, and only while the permission is live.
- The amount stays inside the per-firing limit and the lifetime limit.
- The trigger is true, when the chain can read it.
- The action calls only the venues you signed, and the output goes to your wallet.
- The required result holds.
If any check fails, the whole transaction reverts. Nothing moves, and you pay no fee.
What the agent decides
The agent decides when to fire, how much to spend inside your limits, and the route through the venues you signed. It uses AI to read your instruction, for example "a quarter of my USDC". The contract does not trust these choices; it checks the result of each firing. If the agent cannot work out an amount from your instruction, it does nothing and tells you.
Your tokens stay in your wallet
Your wallet approves the Shield contract to spend the permission's token. Only the Shield can use that approval, and only through a permission you signed. Tokens leave your wallet only during a firing, in one transaction that sends the output to your wallet (for a send, to the one recipient you signed).
Revoke at any time
On the Agents page, delete the agent and choose Revoke and delete: your wallet signs the revocation onchain. Signo does not need to approve it. A permission that is no longer attached to an agent shows in Settings → Wallet permissions, with its own Revoke button. It takes effect when the transaction is included in a block; a firing already included before it still executes. Pausing the agent in Signo stops new firings but leaves the permission live onchain until you revoke it.
Cost and review status
- Signo charges 0.1% (10 basis points) of what each firing spends, on top of your limits. A firing that reverts pays no fee.
- The Shield contracts went through 15 review passes. They have not been formally audited. The code, tests and review history are public in the signo-shield repository.
Get access
New Execute setup is in early access. In the agent builder, choose Execute and press Join waitlist. Notify and Prepare agents are open to everyone and never spend without your signature.
More detail: What you authorize with an Execute agent and Security.