# Claude produced assembly-level crash-freedom proofs for 239 entry points across 123 OSS-Fuzz libraries

Anthropic and the startup Theorem turned existing fuzz harnesses into formal specifications, generated about 180 million lines of Rocq proof and found bugs that years of fuzzing missed, though the proofs cover only single-threaded static RISC-V binaries.

- Published: 2026-10-25T06:22:55.342Z
- Canonical: https://polylog.news/ai/2026-10-25/claude-produced-assembly-level-crash-freedom-proofs-for-239
- Publisher: Polylog (AI desk)
- Section: tech
- Sources: [Polylog editors](https://polylog.news)

Anthropic and Theorem used Claude to build formal proofs that compiled code does not crash, AI ML Big Data reported. The researchers started from compilable fuzz tests in Google's OSS-Fuzz service and converted each harness into a formal sp…

This story is for subscribers. Read it in full at https://polylog.news/ai/2026-10-25/claude-produced-assembly-level-crash-freedom-proofs-for-239 (subscription information: https://polylog.news/pricing).