Anthropic and Theorem formally verify 123 open-source projects with Claude, launch free OSS Verify
Miles_Brundage · x · 2026-10-23
- Anthropic published new research on its Frontier Red Team blog: in partnership with Theorem Labs, Claude was used to formally verify security properties across 100+ real-world open-source libraries in a few weeks.
- Formal verification can prove code is free of certain bugs, but historically only worked for small codebases or required extensive expert time — AI models are changing that.
- Theorem launched OSS Verify, offering free binary-level formal verification to qualifying open-source projects. Unlike tests (covering inputs someone wrote) or fuzzers (inputs they happened to try), a proof covers every input: the theorem targets the compiled binary, not source code, and states no input up to 64 MiB can cause a crash via a failing fuzz test or ASan.
- Projects receive a machine-checked proof or patches for issues found; the goal is bootstrapping a verified software ecosystem.
More from Safety
- 'Fear of Antitrust Suits Won't Slow AI Labs' — They Never Hesitated on Copyright Either — NathanpmYoung · 2026-10-23
- Local AI security checklist: pin Ollama ≥0.35.0, vLLM ≥0.31.0, purge LiteLLM 1.107.0 malware — AIFlow_ML · 2026-10-23
- Fortinet discloses multiple critical and high-severity flaws in FortiOS, FortiManager, FortiAnalyzer — cyb3rops · 2026-10-23
- Malware found in GitHub DaisyUI-MCP repo: luajit.exe trojan steals credentials — AveryFreeman · 2026-10-23
- Naval amplifies take: assume 70% of the Internet was already rooted, AI just exposed it — naval · 2026-10-23
- New paper with 25 coauthors examines feasibility of a hardwired pause on frontier AI training — wfithian · 2026-10-23