How Do Crypto Bridges Work — And Why Are They Such a Hack Target?

Crypto bridges connect blockchains that cannot directly verify one another’s activity. Some move tokens; others pass messages or use liquidity on each side. “Bridge” therefore describes a job, not one security design. The risk depends on the route’s contracts, verification method, administrators, token model, and withdrawal path.

Bridge failures are often grouped together as “bridge hacks,” but that label hides different causes. The Ronin and Nomad incidents show why a useful assessment asks what a specific bridge trusts and how that trust can fail.


What a bridge does

A blockchain normally verifies transactions using its own rules and consensus. It does not automatically know whether an event happened on another chain. A bridge adds a way to communicate that information, or to exchange value across the two systems.

Common asset-transfer patterns include lock and mint, where assets are locked on one chain and a corresponding representation is issued on another; burn and mint, where an asset is destroyed on the source chain and created on the destination; and atomic swaps, where parties exchange assets across chains. Some bridges also carry messages or arbitrary data, not just tokens. A single product may combine mechanisms, so it is inaccurate to assume every bridge locks the original asset and mints a synthetic copy.

Security depends on what verifies the transfer

A bridge has to decide whether a claimed deposit, burn, or message really happened. Designs make different trade-offs:

  • External validators or multisignatures rely on a group of operators to observe events and authorize actions. Security depends on the group’s threshold, independence, key custody, software, and ability to respond to compromise.
  • Light-client or proof-based systems verify information using cryptographic proofs or another chain’s consensus rules. This can reduce reliance on a separate signer set, but it does not remove assumptions about the proof implementation, contract code, upgrades, or the connected chains.
  • Optimistic verification accepts a message unless someone challenges it during a defined period. The challenge rules, available evidence, monitoring, and ability to submit a challenge all matter. There is no universal seven-day period for every bridge using this broad idea.
  • Liquidity networks may pay the user from liquidity already on the destination chain and settle between providers later. That can change the wait, but introduces route, liquidity, provider, and contract risks of its own.

“Trustless” is often used as shorthand for reducing or changing trust assumptions. It does not mean a user can ignore the code, governance keys, verifier design, token representation, or underlying networks.

Two incidents, two different failure paths

Ronin: validator and operational controls

Ronin’s own post-incident account says that 173,600 ETH and 25.5 million USDC were drained from its bridge on March 29, 2022. The FBI later attributed the theft of about $620 million in Ethereum reported that day to Lazarus Group and APT38, actors associated with North Korea.

This case is a reminder that a bridge relying on external authorization depends on more than the smart contract. The authorization threshold, private-key custody, access controls, monitoring, and incident response all affect the outcome. Ronin’s later account describes additional withdrawal limits and human review for its highest-value tier. Those controls illustrate one response to a particular design; they do not establish a guarantee that any bridge using similar controls is safe.

Nomad: a message-authentication bug

Nomad’s August 2022 postmortem describes a flaw in the Replica contract’s message-authentication logic. In an affected setup, the zero value used for an unproven message could also be treated as an acceptable root. That allowed messages that had not been proven to pass the relevant check. Once the exploit became visible on-chain, other addresses copied the pattern.

Nomad said the relevant change was included in an upgrade that went through an audit period before being deployed. The lesson is not that audits have no value. It is that an audit is a review of a defined code and scope at a point in time; initialization, upgrades, configuration, dependencies, and deployed behavior still need careful verification.

Withdrawal time is route-specific

Some rollup bridges use a challenge or finalization period before funds can be withdrawn to the parent chain. Arbitrum’s current user documentation, for example, says withdrawals from Arbitrum One or Arbitrum Nova through its bridge take at least seven days to reach Ethereum mainnet. That timing is specific to those routes. It should not be repeated as a rule for every optimistic bridge or every cross-chain transfer.

A third-party route may offer a faster experience by using liquidity or a different settlement design, but the user is then relying on that route’s own providers and contracts. Compare the exact networks, asset, and exit path rather than relying only on a bridge brand or a “fast” label.

Questions to check before using a bridge

  • Which exact source and destination networks are involved, and is the token native there or a bridged representation?
  • Who or what verifies the source-chain event: the chain’s own proof system, a light client, a validator set, a multisig, or a liquidity provider?
  • Who can upgrade, pause, or change the contracts, and are those actions delayed or publicly observable?
  • What are the normal withdrawal time, fees, liquidity limits, and recovery process for this specific route?
  • What did audits cover, when were they performed, and what code or configuration changed afterward?

There is no single bridge design that removes every risk. The useful comparison is between the exact assumptions and failure modes of the route you plan to use. A bridge that is convenient for one asset and network pair may have different security and withdrawal properties on another.

Sources

See Also

Related Terms