Address poisoning is a scam in which an attacker plants a look-alike wallet address in a victim’s transaction history, hoping the victim later copies the fake address instead of the real one and sends funds to the attacker. In the field of crypto security & rules, it is classed as a social engineering attack on user habits rather than a technical exploit: no private key is stolen, no smart contract is broken, and every transaction the victim signs is valid on the blockchain.
The attack works because blockchain addresses are long hexadecimal or Base58 strings that people rarely read in full. Most users check only the first and last few characters, and many copy a recipient from recent activity rather than from a verified address book. Address poisoning exploits exactly that shortcut, and because on-chain transfers are irreversible by design, a single careless paste can cost the entire amount sent.
How an address poisoning attack works
An address poisoning attack follows a predictable four-stage pattern: the attacker monitors the chain, generates a matching address, seeds the victim’s history and waits for a mistaken transfer. The scheme is cheap to run at scale, which is why it targets thousands of wallets at once.
- Monitoring: automated bots watch public ledgers for wallets that regularly move meaningful amounts, especially stablecoins such as USDT or USDC.
- Vanity generation: the attacker uses GPU-based address generators to create a vanity address whose first and last characters match a counterparty the victim already uses.
- Seeding the history: the attacker sends a zero-value transfer, a dust amount or a fake token transfer so the look-alike address appears in the victim’s recent activity.
- Waiting for the paste: when the victim next pays that counterparty and copies the address from history or a block explorer, the funds go to the attacker.
On EVM chains such as Ethereum, a common variant abuses the ERC-20 transferFrom function, which allows zero-amount transfers to be emitted on behalf of any address without that owner’s signature. The result is a transaction that appears to come from the victim’s own wallet, making the poisoned entry look like the victim’s own past activity.
The following table summarizes the main seeding techniques and why each one fools users.
| Technique | What appears in history | Why it deceives |
|---|---|---|
| Zero-value transfer | A 0-token transfer from the victim | Looks like the victim’s own outgoing payment |
| Dust transfer | A tiny real amount from the look-alike | Looks like an incoming test payment |
| Fake token transfer | A counterfeit token with a real ticker | Mimics a familiar stablecoin movement |
| Mirrored amount | A fake token for the same amount just sent | Matches the size of a recent genuine payment |
Why look-alike addresses are convincing
Look-alike addresses are convincing because wallet interfaces truncate addresses and humans compare only fragments. An Ethereum address has 40 hexadecimal characters, yet many wallets show something like 0x1a2b...9f8e, leaving only about eight visible characters to check.
Matching more characters costs the attacker exponentially more computation. Each extra hexadecimal character multiplies the search effort by about 16, so attackers usually match four to six characters at each end. That is enough to defeat a quick glance but not a careful full comparison, which is the core weakness defenders should target.
The table below shows how the number of matched characters relates to attacker effort and the likelihood of detection by an attentive user.
| Characters matched at each end | Relative attacker effort | Chance a quick glance detects it |
|---|---|---|
| 3 + 3 | Trivial, seconds on a laptop | Moderate |
| 4 + 4 | Low, minutes on a GPU | Low |
| 5 + 5 | Moderate, hours on GPU clusters | Very low |
| Full 40 characters | Computationally infeasible | Not applicable |
Address poisoning compared with related crypto scams
Address poisoning differs from related scams because the victim signs a correct transaction to a wrong recipient, without granting any permission or revealing any secret. Wallet drainers and approval phishing, by contrast, trick the user into signing a permission that lets the attacker move funds directly.
The comparison below helps place the attack among neighboring threats in the same Topic.
| Scam | What the victim does | What the attacker gains |
|---|---|---|
| Address poisoning | Sends a payment to a look-alike address | The single amount sent |
| Approval phishing | Signs a token allowance for a malicious spender | Ongoing access to approved tokens |
| Wallet drainer | Connects to a fake site and signs malicious requests | Many assets in one sweep |
| Clipboard hijacker | Pastes an address altered by local malware | The amount sent, on an infected device |
| Pig butchering scam | Deposits repeatedly into a fake platform | Large sums over weeks or months |
A clipboard hijacker is the closest technical cousin, but it requires malware on the victim’s device. Address poisoning needs no device compromise at all, which is why it scales across entire chains and why antivirus software offers no protection against it.
Detecting and preventing poisoned transfers
Address poisoning is prevented mainly by never copying a recipient from transaction history and by verifying the full address through an independent channel. Technical controls in wallets and explorers help, but the decisive defense is a disciplined payment routine.
The decision table below matches common payment situations with a recommended practice.
| Situation | Recommended practice | Reason |
|---|---|---|
| Recurring payment to a known party | Use a saved address book entry | Removes reliance on history |
| First payment to a new counterparty | Send a small test transfer and confirm receipt out of band | Limits loss if the address is wrong |
| Large treasury transfer | Require multi-person verification of the full address | Adds an independent check |
| Unexpected zero-value entries in history | Treat them as poisoning attempts and ignore them | They are a known warning sign |
| Human-readable name available | Resolve and cross-check the name service record | Reduces character-by-character comparison |
Many wallets and block explorers now hide or flag zero-value transfers and spam tokens, and some display a warning when the recipient resembles a past counterparty. These filters reduce exposure but are not complete, since attackers adapt by switching to dust amounts or new token contracts.
Limitations, recovery and regulatory context
Recovery after address poisoning is rarely possible, because blockchain transactions cannot be reversed and attackers typically move stolen funds through swaps and mixers within minutes. The realistic outcomes are partial and depend on speed.
- Stablecoin freezes: issuers of centralized stablecoins can sometimes freeze the attacker’s address if notified quickly.
- Exchange cooperation: if funds reach a regulated crypto exchange, KYC and AML controls may allow the account to be frozen and identified.
- Sanctions screening and tracing: analytics firms label poisoning addresses, which helps compliance teams block onward flows.
- Negotiation: in rare high-value cases, attackers have returned funds after public on-chain messages, but this is not a strategy to rely on.
From a regulatory perspective, address poisoning is treated as fraud or theft under general criminal law. Frameworks such as MiCA and the travel rule do not prevent it directly, yet they strengthen the points where stolen funds meet regulated service providers, which is where most successful interventions happen.
In simple terms
Imagine your phone shows a list of people you recently paid. A con artist quietly adds a fake look-alike contact to that list with a name that looks almost identical to your landlord’s. Next month you scroll, tap the familiar-looking entry and pay the wrong person.
Address poisoning does exactly this with crypto addresses. The fake entry costs the scammer almost nothing, and the money you send cannot be pulled back. The simple fix is to never pick a recipient from your history and to check the whole address, not just the start and the end.
A practical example
A small trading company pays a supplier about 20,000 USDT every week from the same Ethereum wallet. An hour after one routine payment, a zero-value transfer appears in the wallet history from an address that starts and ends with the same five characters as the supplier’s address.
The following week, a new finance assistant opens the block explorer, copies the most recent outgoing address and prepares the payment. Because the company had adopted a policy requiring a second colleague to compare the full address against the saved address book, the reviewer notices that the middle characters differ and stops the transfer.
The company then adds the supplier to the wallet’s address book, enables spam filtering of zero-value transfers and adds the incident to staff training. No funds are lost, and the near miss becomes the reason the team treats every unexpected history entry as a warning sign.