Skip to content

Free and read-only. Nothing here can move money.

Labs · Agent Wallet Watchdog

Know within minutes when your agent's wallet drains, loops or pays someone new. Without watching a dashboard.

Paste the agent's public address. See what it did in the last few hours, then leave it with us: we learn its normal from its own history and check it every few minutes on Base, Ethereum, Arbitrum, OP Mainnet and Polygon. You hear about it by email or Telegram.

Read-only. We never ask for keys, can't sign anything and can't move funds.

1. Scan a wallet now

Chains

On the free path we follow the tokens listed next to each chain, plus every Permit2 approval.

Want a second pair of eyes on one real workflow?

This tool checks what an agent wallet did onchain and keeps telling you when it does something it shouldn't from what you typed. It can't see your code, keys, contracts or production traffic. If the result shows gaps, send me one real workflow. In my free 48-hour integration review I go through it end to end and send you a written report: fit, dependencies, permission risks, acceptance tests and a build or don't-build-yet call.

I look at how your agent holds and spends money: where the caps live, who can approve what, and what should stop the agent automatically when one of these alerts fires.

Request the free 48-hour review (opens Cal.com in a new tab)

Contact: Shilpa Goel, [email protected]

This is an engineering checklist, not a security audit, legal, compliance or investment advice.

Why an alert has to watch for you

Founders keep telling me their spend control comes down to "set an alert and hope someone's watching", while an agent in a tight loop can burn hundreds of dollars in minutes. A chat assistant can explain a transaction you paste. It can't sit on your wallet at 3 a.m. This does: it reads the chain on a schedule, remembers what normal looks like for your agent, and writes only when something breaks the pattern.

Shilpa Goel, Consistra

What it watches

  • Drains and floors: a balance falling fast, or below the level the agent needs.
  • Approvals: unlimited or long-lived allowances (ERC-20 and Permit2), and any approval to a plain wallet, the drainer pattern.
  • Loops and spikes: many payments in minutes, or an hour far above this wallet's typical hour.
  • Paid twice: the same amount to the same payee again within minutes.
  • New payees and odd hours, learned from the wallet's own past, plus sanctioned counterparties, always.

Coverage

  • Up to 3 wallets per inbox or Telegram chat, checked every 5 minutes
  • USDC on every chain, plus USDT, EURC, DAI and wrapped ETH or BTC where they are common (listed per chain), ETH, and every Permit2 approval
  • Email or Telegram, 30 days of alert history

What this tool can't see

  • Native ETH or POL sent straight from the wallet shows up as a balance change, not as a payment with a payee, so it feeds the drain rule only.
  • We follow a fixed list of contracts per chain. Other tokens and NFT approvals are not covered yet.
  • Free public endpoints keep limited history (about a day on Ethereum). Ranges we couldn't read are listed as gaps, never counted as quiet.
  • Solana and other non-EVM chains are not covered yet.
  • An alert is a tripwire, not a limit. Caps the agent can't pass have to live in the wallet or signer policy.

Free starter checks that go with it