Direct answer

The risks specific to a DEX are: impersonator tokens (anyone can list), honeypot contracts (buy permitted, sell blocked), malicious approvals (permission granted, not keys stolen), thin pools (impact is proportional to size relative to reserves), cloned interfaces, and MEV extraction from your pending transaction. Each has a specific check, and together they take a few minutes.

Risk 1 — Impersonator tokens

Why it exists here. Listing is permissionless. Nobody reviewed anything, and names and tickers are not unique — anyone can deploy a contract calling itself by an established project's name and create a pool for it.

The check. Get the contract address from the project's own domain or verified account, and search by address, not by name. Confirm the chain, since the same project may deploy on several with different addresses.

Tell: a token that appears in search results and social replies around an announcement, promoted with an address that did not come from the project.

Risk 2 — Honeypot contracts

Why it exists here. No contract review stands between the deployer and a live market.

The check. Before buying, look at the contract: is source verified? Does it have a blacklist, pause, or an asymmetric transfer fee — higher on sells than buys, or one the owner can raise later?

Tell: the price rising steadily while you find no evidence of anyone selling. On a honeypot, rising is the expected behaviour.

Full routine: how to verify a smart contract.

Risk 3 — Malicious approvals

Why it exists here. Trading requires granting contracts permission to move your tokens, and unlimited approvals are the interface default.

The check. Read every approval: which token, which contract, what amount. Prefer finite. Treat a permit — gasless, signature-only — with the same care, because it grants identical authority while feeling inconsequential.

Tell: any prompt requesting an approval when the action you initiated should not need one.

The structural defence: trade from a wallet holding almost nothing. This works without requiring you to identify malice correctly.

Risk 4 — Thin pools

Why it exists here. A constant-product pool never runs out of liquidity. It fills any order, at a progressively worse price.

The check. Look at pool reserves and the displayed price impact before confirming. Execution impact is your trade size divided by the pool's quote reserve — so it is your size relative to the pool that matters, not its absolute size.

Tell: double-digit price impact on an ordinary trade size. The trade will still succeed, which is precisely the trap.

Risk 5 — Cloned interfaces

Why it exists here. The front end is a normal website on a normal domain, whatever the contracts behind it are.

The check. Type the domain yourself or use a bookmark you created that way. Never a search advertisement, a message, or a reply.

Tell: you cannot reliably detect a clone by looking, which is why the defence is procedural rather than perceptual. A password manager helps — it will not autofill on the wrong domain.

Risk 6 — MEV and sandwich attacks

Why it exists here. Pending transactions are public, and AMM pricing is deterministic — so the effect of your trade is computable by anyone before it confirms.

The check. Keep slippage tolerance at the smallest value that reliably executes, since your tolerance is the budget available to a sandwicher. Split large trades. Prefer deep pools. Use a private transaction relay where available.

Tell: consistently filling near the worst edge of your tolerance.

Mechanism: how decentralized exchanges work.

Risk 7 — Rug pulls

Why it exists here. Nothing prevents the party who supplied the pool from removing it.

The check. Is liquidity locked, verifiably on-chain, with a meaningful expiry? Who supplied it? What is the holder concentration?

Tell: a claimed lock nobody can point to on-chain. An unverifiable lock is no lock.

Full routine: how to spot a rug pull.

Risk 8 — Contract risk in the protocol itself

Why it exists here. Even a well-audited protocol can contain an exploitable flaw, and upgradeable contracts can change after you inspect them.

The check. Prefer protocols with a long live track record and substantial value at stake. Understand that an audit is evidence someone competent looked, not a guarantee.

Tell: none reliably available in advance. This is the residual risk that diligence reduces but does not remove — which is an argument about position size, not about checking harder.

The combined routine

Contract address from a project-controlled source; searched by address
Chain confirmed
Source verified; no asymmetric or mutable sell fee
Liquidity lock confirmed on-chain, if claimed
Holder concentration acceptable
Pool reserves checked; price impact acceptable at my size
Site reached by typing the domain
Slippage tolerance minimal
Trading from a wallet sized to this trade
Every signature read; approvals finite
Exit path considered

The two habits that do the most

Trade from a wallet holding only what you are trading. Bounds every mistake.
Verify identifiers, not labels. Contract addresses over token names; typed domains over links.

Neither requires you to correctly judge whether something is malicious — which matters, because that judgement is what attackers spend their effort defeating.

How to Spot Common DEX Risks Before They Cost You