On Code4rena's 2024-07 Karak restaking contest, our AI engine reproduced the major contest findings AND surfaced 3 additional HIGH-severity issues not in the contest's published HIGH/MEDIUM report — including a single-transaction operator rug — all verified with runnable Foundry PoCs.
AI smart contract audit engine caught every HIGH finding on Code4rena veRWA, plus an additional division-by-zero issue not in the contest's published HIGH/MEDIUM report.
Our AI smart contract audit engine caught all 7 HIGH findings on Code4rena BakerFi, plus 15 of 16 MEDIUMs, including the EIP-2612 permit-signature replay in VaultRouter.
Our AI Solana audit engine caught 100% of Critical and 90% of HIGH findings on Jito Restaking — 9k lines of Rust across four prior audits.
2/2 Critical and 6/6 High vulnerabilities reproduced on OWASP Juice Shop. 90.3% OWASP Top 10 coverage. Real numbers, real targets, no cherry-picking — measured against a public ground truth catalog.
Our AI smart contract audit engine reproduced every HIGH and MEDIUM finding from Code4rena's VTVL contest report, with a passing Foundry PoC for each.
Our AI smart contract audit engine caught every HIGH-severity finding on Code4rena Wildcat — 6/6 HIGH and 8/10 MEDIUM, scored against the official contest report.
Our AI smart contract audit engine solved all 7 Ethernaut + Damn Vulnerable DeFi challenges — reentrancy, flash-loan, share inflation, gas DoS.