What is the Liquid Security Index?
A public security score from 300 to 1000 for DeFi protocols, tokens, companies and prediction markets.
Higher is safer. BlackHart-1 reads the code on every release and rescores it, so you can compare risk before you deposit, trade or sign.
Aave V3ProtocolMithril886
LidoProtocolDamascus847
USDCTokenTempered733
Coinbase / BaseCompanyTempered730
PolymarketProtocolForged646
Public scores run 7 days behind. See all 110
Why it matters
Crypto loses money to exploits and failures. People deposit, trade and sign every day with no common way to compare the risk.
The index gives every protocol a score from the same method, in public, so the risk is visible before the money moves.
See the hacks feedHow it works
BlackHart-1 is the model behind the index. It reads a protocol's code on every release and scores it across 12 dimensions, each with a published weight.
Code ships
A release, an upgrade or a new parameter goes live.
BlackHart-1 scores
It reads the new code and scores all 12 dimensions again.
The score moves
Only when the code or the method changes. Every move is labeled Release or Model update.
849847Model updateLido, May 20
The 12 dimensions and their weights
- Access control18%
- Economic soundness13%
- Oracle integrity13%
- Compositional risk5%
- Governance10%
- Maturity12%
- Adversarial resilience10%
- Supply chain4%
- Liquidity and market structureWhere it applies
- Cross-chain messagingWhere it applies
- Operational security10%
- Cascade exposure5%
A prediction market has no code to read, so it gets 6 dimensions of its own.
See the full methodWhat the numbers mean
Every score falls in one of 7 forge tiers. A high score means the code holds up well against attack. A low score flags where to look harder before you deposit, trade or sign.
- Adamantine950 to 1000
- Mithril850 to 949
- Damascus750 to 849
- Tempered650 to 749
- Forged550 to 649
- Cast450 to 549
- Raw300 to 449
What the score saw
69 exploits tracked since April 2026. 15 hit protocols we had scored before the exploit. Here is what the score saw first.
- Across$350K lost776Score before the exploit
Our pre-hack assessment of Across focused on its contracts and code practices and did not measure domain or registrar security, so this attack surface was missed. The contracts it did assess were not touched.
Read the forensics - Polymarket$2.7M lost646Score before the exploit
Our assessment of Polymarket was not rescored after the June vendor attack, and its supply-chain review still looked only at smart-contract dependencies, so third-party website scripts stayed outside what we measured. It did rate operational security as the weakest area, which is where front-end script controls belong.
Read the forensics - Swell$6K lost732Score before the exploit
Our pre-hack assessment of Swell flagged operational security as its weakest area and noted that its exchange rates come from Swell's own oracle, but it could not see how many hot keys shared one compromise path. It also flagged the upgradeable contracts, the path the attacker went for, and there the multisig and timelock held.
Read the forensics
Every move is public
Every protocol page has a history chart. Each move shows its date, the score before and after, and whether it was a release or a model update.
- Model update849 to 847down 2May 20
- Model update847 to 849up 2May 19
- Model update843 to 847up 4May 19
Who it's for
People with money on the line
Check a score before you deposit, trade or sign. Search any protocol on the site, or add the browser extension for scores on X, Polymarket and DefiLlama, and a check in your wallet before you sign.
Protocol teams
Builder gives you a private score on every push, before it goes live, with vulnerabilities by severity and file. $99 a month per repo.
Integrators
Put the score on your site, README or app with one snippet: a card, a badge or the API. Free with a 7-day delay and attribution.
Free, or Live
Every public score is free and runs 7 days behind. Live shows every move the moment it lands.
Public
FreeEvery score, 7 days behind.
- Protocol scores 7 days behind, market scores 24 hours behind
- Search and every protocol page
- Free chicklets with attribution
Live
$29 a monthReal-time scores and email alerts for your watchlist.
- Real-time scores for protocols, tokens and prediction markets
- Email alerts for every protocol you watch
- Every Score Move as it lands
Questions
- Is this an audit?
- No. An audit reviews the code once. BlackHart-1 rescores the code every time it changes, and the audits a protocol has had count as one input.
- Who decides the score?
- BlackHart-1 and the published method. A score combines 12 dimensions, each with a published weight. Commercial relationships do not influence it.
- Can a protocol pay for a better score?
- No. Every protocol, paying or not, is scored with the same 12 dimensions, the same tooling and the same standard of evidence. Paid plans add speed and private detail, never a better score.
- How often does it change?
- Whenever the code or the method changes. BlackHart-1 rescores a protocol when its code changes, and every move is labeled Release or Model update. Prediction markets are rescored daily, and again whenever their trading changes sharply.
- What if a score is wrong?
- Any protocol can submit evidence of a factual error for review, at no cost. Every score can be traced back to the observable on-chain behavior that produced it.
Check the score before you sign.
Every public score is free. Live shows every move as it happens.