The Provable Consensus Client — a formally verified @leanEthereum client built with @leanprover
Joined April 2026
- Tweets36
- Following11
- Followers139
- Likes23
⚠ the calendar timeline has also shifted later, so this does not necessarily mean earlier delivery
↓ previous version
The strawmap has been updated!
- Shorter slots and PQ upgrades move to earlier forks
- leanVM frames added to the EL roadmap
- Uncertainty marker removed from “leanCL STF in Lean 4”
@leanEthereum
strawmap.org/
We get step to next stage 🧙
to prove the correspondence of spec and implementation with Aeneas
github.com/NyxFoundation/ver…
For examaple, We are prototyping Option 4 inspired by evm-asm✨
github.com/NyxFoundation/cl-…
@verityclient is currently using Option 2, but with the discovery of several alternative approaches, it is considering a change in architecture.
@verityclient is currently using Option 2, but with the discovery of several alternative approaches, it is considering a change in architecture.
“Formally verified client” doesn’t automatically mean the binary users run is verified.
This article explores where proofs end, what remains in the TCB, and five paths toward E2E verification
ethresear.ch/t/ethereums-tcb…
We experimentally integrated this into Verity through a C ABI
github.com/NyxFoundation/ver…
We just merged Lean4 implementation of SSZ!
Still a huge amount to iterate on, especially making the Astra-generated proofs more human-readable.
But Lean/Python cross-testing in CI has already caught a few subtle bugs in the Python spec :)
github.com/ethereum/ssz-spec…
2/10
Formal verification mathematically proves that both the specification and the code behave correctly.
With @leanprover we can catch safety issues that ordinary testing often misses.
lean-lang.org/
1/10
Do you know Decoupled Consensus?
It resolves multiple problems long faced by @ethereum ’s consensus algorithm.
We are formally verifying it in @leanprover
ethresear.ch/t/unblocking-fa…
