QEDGen

Formally verify Solana programs with Lean 4 proofs. An agent skill powered by Mistral's Leanstral model.

18
1
100% credibility
Found Mar 29, 2026 at 18 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Rust
AI Summary

A tool and agent skill for generating AI-assisted formal mathematical proofs to verify the correctness of Solana smart contracts.

How It Works

1
πŸ” Discover QEDGen

You hear about a helpful tool that can mathematically prove your Solana smart contract code is correct and safe from bugs.

2
πŸ“₯ Get it set up

Run a quick installer script that prepares everything on your computer, including downloading what's needed.

3
πŸ”— Link your AI helper

Connect to a free AI service so the tool can use smart thinking to create proofs.

4
πŸ“„ Share your program details

Point the tool to a simple description file of your Solana program.

5
✨ Watch proofs appear

The tool automatically generates mathematical proofs checking things like access rules, money transfers, and safety limits.

6
βœ… Verify the proofs

Run a quick check to ensure the proofs compile perfectly with no gaps.

πŸŽ‰ Proof complete!

Celebrate as your Solana program now has unbreakable mathematical guarantees it's correct and secure.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 18 to 18 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is solana-skills?

solana-skills lets you formally verify Solana programs with Lean 4 proofs, powered by Mistral's Leanstral model. Feed it your program's IDL, and it generates proofs checking access control, CPI parameters, state machine lifecycles, and arithmetic safety against overflows. Built as a Rust CLI and agent skill, it integrates with tools like Claude or Cursor via `npx skills add qedgen/solana-skills`, outputting compilable Lean code you verify with `lake build`.

Why is it gaining traction?

Formal verification for Solana is niche, but this automates Lean proofs like those for Kyber encryption or TLS 1.3 handshakes, adapted for blockchain patterns such as escrow state transitions or token CPI calls. Devs love the iterative agent workflow: generate SPEC.md, proofs, fill hard subgoals with `qedgen fill-sorry`, and consolidate into one project. It bridges AI agents and theorem proving, slashing manual Lean work.

Who should use this?

Solana program authors auditing escrows, DEXes, or permissioned vaults, especially those chasing audit badges or handling millions in TVL. Anchor devs wanting proofs beyond fuzzing, or teams inspired by rareskills solana skills for zero-knowledge circuits and arithmetic chisel designs.

Verdict

Grab it if you're serious about Solana proofsβ€”early examples like escrow and percolator work out of the box with a free Mistral key. At 18 stars and 1.0% credibility, it's raw and needs more battle-tested programs, but the Leanstral integration makes formal verification feasible today.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.