Lean v4.35.0 Adds Cross-Checking With Verified Kernels
Lean v4.35.0 introduces the `lake check` command, which exports a development and rechecks it with the official `leanchecker` in a sandbox. The release adds the nanoda, lean4lean, con-leche and con-ron kernels. Its `--paranoid` mode runs all four alongside a variant of the official kernel built with a hardened memory allocator, which converts memory corruption into a checked abort. According to the Lean FRO, Claude (Fable and Opus), under Joachim Breitner’s supervision, implemented con-leche and its consistency proof. AI was also used to translate con-leche into Rust, producing con-ron, whose refinement of con-leche’s checking algorithm was proved with Aeneas. The Lean FRO expected lean4lean’s verification to be completed in the following months.
The release gives Lean developers a built-in way to recheck developments with multiple external kernels and compare their results. The hardened official-kernel variant adds protection against memory corruption, while the Lean FRO says checking with both con-leche and con-ron offers substantial protection against implementation bugs shared by Lean’s compiler and runtime. The report also identifies scalability as a major remaining limitation because the current external-kernel approach does not scale to the largest developments.
2026-10-15
AI agents are very effective at finding vulnerabilities in our software. This includes the kernel (the proof checker) of the Lean proof assistant. See our two recent postmortems (August 1, August 24). We are strengthening the official kernel, adding new kernels, and building new tools that simplify rechecking your Lean developments with external kernels. Lean v4.35.0 comes with the lake check command. This command exports your development and rechecks it with leanchecker, the official Lean kernel without all the baggage around it. The whole process runs inside a sandbox. You can disable the sandbox using lake check --inadvisably-no-sandbox. Lean now ships with four additional kernels: nanoda, lean4lean, con-leche, and con-ron. You can recheck your development with all of them using the command lake check --paranoid. In this mode, we also run a variant of the official kernel built with a hardened memory allocator, which turns memory corruption into a checked abort instead of a silently wrong answer.
Con-leche is the first kernel for Lean written in Lean and proven consistent. It was implemented and proven consistent by Claude (Fable and Opus), under heavy supervision by Joachim Breitner at the Lean FRO. "Consistent" here means that every environment con-leche accepts has a model in set theory, following Mario Carneiro's model of Lean's type theory. The proof idea is written up here. The proof itself is checked by the official Lean kernel, and it assumes the consistency of the set theory used in the model. This approach eliminates a complete class of bugs: programming mistakes in the kernel implementation. It does so without a performance penalty. On the full Mathlib export, con-leche executes 0.69x the instructions of the official kernel and checks the whole library in about 100 seconds using eight threads (see PERF.md).
#Lean v4.35.0#Lean FRO#Claude#Aeneas#Mathlib#nanoda#lean4lean#con-leche#con-ron