Smart-contract risk, bridges and hacks

What you will learn

• The main ways smart contracts fail: code bugs, admin keys, oracles and upgrades.
• How cross-chain bridges work and why they are large targets.
• What three documented bridge hacks teach about different failure types.
• How to combine several risk layers into one rough estimate of expected loss.

Explanation

A smart contract is a program stored on a blockchain that holds and moves tokens by fixed rules. Scams and phishing were covered in the Bitcoin course. This lesson is about a different danger: the protocol is honest, but the code or its setup fails.

Main failure types

• Code bugs: for example reentrancy, where a contract sends funds before updating its records, so an attacker can call it again and withdraw repeatedly. The DAO on Ethereum lost about 3.6 million ETH this way in 2016.
• Admin keys: many protocols have privileged keys that can pause, change or upgrade contracts. If a key is stolen or misused, users lose funds even when the code is correct. A multisig (multi-signature wallet) needs several keys, for example 5 of 9, to approve an action.
• Upgradeable contracts: code can be replaced. A timelock delays upgrades, for example by 48 hours, so users can leave first.
• Oracle manipulation: an oracle feeds external prices to a contract. If it reads a thin market, an attacker can move the price with a flash loan (a loan borrowed and repaid in one transaction) and trick a lending protocol.

An audit is a review of code by a security firm. It reduces risk but does not remove it; several hacked protocols were audited.

Bridges

A bridge moves value between blockchains. The common "lock-and-mint" design locks tokens in a contract on chain A and mints a wrapped token on chain B. The wrapped token is only an IOU. If the locked tokens are stolen, or unbacked tokens are minted, the wrapped token can lose most of its value. Bridges hold large pools in one place, which makes them attractive targets.

The lesson: risk sits in code, in key management and in upgrades. In the Wormhole case a related company replaced the funds; in other cases users were not fully repaid. Repayment is never guaranteed.

Worked example

Hypothetical example: you deposit €5,000 of a bridged stablecoin into a lending protocol on a Layer 2 network and expect 6% per year. You assign rough, invented annual failure probabilities to each layer: bridge 2%, Layer 2 network 1%, lending protocol 3%, price oracle 1%. Assumptions: layers fail independently; any failure means total loss; fees, gas and taxes ignored.

1. Probability that no layer fails = 0.98 × 0.99 × 0.97 × 0.99 ≈ 0.932.
2. Probability of at least one failure ≈ 1 − 0.932 = 0.068, about 6.8%.
3. Expected loss ≈ €5,000 × 0.068 ≈ €342.
4. Expected yield = €5,000 × 6% = €300.
5. Risk-adjusted result ≈ €300 − €342 ≈ −€42 per year.

Compare: holding the stablecoin on its native chain in a simpler protocol with 1% and 2% risks gives 1 − (0.99 × 0.98) ≈ 3.0%, an expected loss of about €149 against perhaps 3% (€150) yield. Each extra layer must pay for itself.

Common mistake

Practice activity

On paper: you consider depositing €10,000 in a pool that promises 9% per year. The tokens arrive via a bridge (assume 3% annual failure risk), the chain itself 2%, and the pool contract 4%. Assume independence and total loss on any failure. Calculate the probability of at least one failure, the expected loss and the risk-adjusted result.

Key takeaways

• Smart-contract risk includes code bugs, admin keys, upgrades and oracles.
• Wrapped tokens are IOUs; if the bridge fails, they can become nearly worthless.
• Each layer you add (bridge, network, protocol, oracle) multiplies the chance of failure.
• Audits and timelocks reduce risk; they never remove it, and repayment after a hack is not guaranteed.

Next step

The next lesson looks at NFTs and the tokenisation of real-world assets: what the token actually gives you and where the risks are.