Anthropic uses Claude to formally verify security properties in 100+ open-source libraries

4 people · 6 posts · first post Thu 22:01 UTC · latest 2 days ago

Anthropic and Theorem report using Claude to produce assembly-level formal proofs of memory safety across 123 real-world open-source libraries, finding bugs that survived trillions of test runs. They frame it as a defense against AI-enabled exploit chaining. Theorem launched OSS Verify offering free proofs to open-source projects; Miles Brundage praised the talent joining Theorem.

All posts · 2, most important first

1. @anthropicai org · Thu 22:01 UTC
   New on our Frontier Red Team blog: using Claude to formally verify software at large scale. Formal verification can prove that code is free of certain bugs, but historically it has only been feasible for small codebases (or required a lot of expert time). With AI models, however, this is beginning to change. In collaboration with @theoremlabs, we used Claude to prove security properties in over one hundred real-world open-source libraries over the span of a few weeks. Read our research in full here: anthropic.com/research/forma…
   817 likes · 104 reposts
2. @miles_brundage policy · Thu 23:04 UTC
   Lots of very sharp folks seem to have joined Theorem lately Not surprising if you've met @rajashree but still cool to see!
   1 likes · 0 reposts

Reposted by @hamandcheese, @jackclarksf, @miles_brundage.