What it is
Bridges have lost more user funds than any other crypto primitive — billions across Ronin, Wormhole, Nomad, Multichain, and others. Yet users cross them daily on vibes alone. This tool scores any bridge on the four factors that actually predict disasters: custody model, size of the honeypot, audit freshness, and who holds the keys. Answer four questions, get a 0–100 risk score plus the exact checklist professionals run.
It won’t bless a bridge as safe — nothing can. It will tell you, in plain language, whether you’re crossing a rope bridge or a highway, and what to do differently in each case.
Why decentralization matters
Every bridge is a decentralization downgrade you accept voluntarily: assets leave chains with thousands of validators and enter contracts guarded by committees, multisigs, or single companies. The ecosystem’s biggest thefts all exploited this exact downgrade. Scoring bridges honestly — and preferring native paths (rollup official bridges, IBC) — is how users vote with their funds for architectures that don’t require trusting five strangers with billions.
How the formula works
Four weighted inputs produce a 0–100 risk score (higher = riskier): bridge type (custodial multisig 35 / optimistic or ZK 15 / native rollup bridge or IBC 8, weight ~40%), audit status (none +25 / old +12 / fresh 0, weight ~25%), honeypot size by TVL (>$1B +20 / $100M–1B +10 / below +4, weight ~20%), and key distribution (single entity +15 / small multisig +8 / decentralized +0, weight ~15%). Weights reflect post-mortem analysis: custody model dominates because it has dominated every major loss.
Calculator
Risks
Scores model known failure modes, not unknown ones — novel cryptography, governance attacks, and slow multisig rot escape any questionnaire. A low score is permission to proceed carefully, never a safety certificate. Size every crossing by what you’d survive losing, because bridges have repeatedly proven that even “safe” ones fail.
Worked example
Score a typical third-party bridge: custodial multisig (35) + $800M TVL (10) + audit from 2022 (12) + 5-of-9 signers (8) = 65 raw → normalized ~68/100, HIGH RISK verdict. Now the same route via a native rollup bridge: proof-based (15) + same TVL (10) + fresh audit (0) + decentralized (0) = 25 → ~26/100, LOWER RISK. The 40-point gap is almost entirely custody model and audit freshness — which tells you exactly where to focus diligence: who holds the keys, and who checked last.
Common mistakes
Grading the bridge you *want* to use instead of the one that exists (optimism about unaudited code is how post-mortems start). Counting “backed by big VCs” as security (investors don’t co-sign your refund). Ignoring the withdrawal path until you need it urgently — discover challenge windows and liquidity crunches in advance, not mid-crisis. And reusing one bridge for everything, concentrating personal exposure exactly like the protocols shouldn’t.
FAQ
Can any bridge score zero? No — minimums reflect irreducible risks (smart-contract bugs, black swans). Trustlessness is asymptotic, never achieved.
TVL: safer big or small? Big TVL means battle-tested but juicier target; small means untested. Mid-size with fresh audits and modest incentives is often the sweet spot.
Official vs third-party? Official (native) bridges inherit the L1/L2 security model; third-party adds a whole company plus its contracts to your trust surface. Prefer native unless speed demands otherwise.
Related
- Bridge · Transaction · Blockchain
- Bridges and Cross-Chain · Sending and Receiving Like a Pro
- Arbitrum audit — L2 security done with Ethereum anchoring.