TECHNICAL REPORT — TL;DR
Every AI audit tool launching in 2026 has the same hidden weakness.
They reason about code without grounding in actual audit precedent. No retrieval substrate underneath.
SCAR is that substrate.
Retrieving over 232,000 Ethereum smart contracts: BM25 collapses. R@10 = 0.31
SCAR holds. R@10 = 0.90
Same hardware SCAR is 47× faster (114ms vs 5,433ms).
The missing layer underneath every audit AI👇
Smart contract security gets stronger when past findings are easy to reuse.
@ScarRetreival by @LifeHodlr is an open-source AI retrieval system that helps auditors and AI tools quickly surface relevant vulnerability precedents from 232,000 real Ethereum contracts.
Learn more: https://t.co/yADATrvBzf
Support: https://t.co/DvPNJUG1dW
On top of being more accurate, Scar is also faster than dense retrieval and BM25 at full-corpus scale.
— 6× faster end-to-end vs Dense (114 vs 725ms P50)
— 2.3× smaller index than Dense (299 vs 678 MB)
— ~50× faster than BM25's Python implementation
inverted-index compatible
exact top-k no ANN approximation.
TECHNICAL REPORT — TL;DR
Every AI audit tool launching in 2026 has the same hidden weakness.
They reason about code without grounding in actual audit precedent. No retrieval substrate underneath.
SCAR is that substrate.
Retrieving over 232,000 Ethereum smart contracts: BM25 collapses. R@10 = 0.31
SCAR holds. R@10 = 0.90
Same hardware SCAR is 47× faster (114ms vs 5,433ms).
The missing layer underneath every audit AI👇
SCAR is in @Giveth 's Ethereum Security QF round, closing May 15.
Funded by TheDAO Security Fund — repurposing 75k ETH from the 2016 hack into ethereum security work.
QF math: $5 from 100 unique wallets matches more than $500 from one whale.
https://t.co/IQu3fsL23f
"I remember reading this exact bug 18 months ago in a @code4rena contest. I cannot, for the life of me find it again."
Auditors don't have a search problem. they have a recall problem.
Scar fixes that.
R@10=0.901 over 232k smart contracts at 114ms.
live demo at https://t.co/Kfr142IDAi.