What is the Aristotle Lean skill?

Aristotle Lean is a specialized Claude Code capability that leverages the Harmonic API to provide gold-medal level theorem proving and formal verification for Lean4 projects.

How does Aristotle handle incomplete proofs?

The skill identifies 'sorry' markers within Lean4 code and uses advanced reasoning models to automatically generate valid proof steps to fill those gaps.

What are the performance benchmarks for this skill?

Aristotle has achieved a 90% score on the MiniF2F benchmark and 96.8% on the VERINA benchmark, placing it at a world-class level for automated mathematical reasoning.

Does it support external libraries like Mathlib?

Absolutely. It is fully integrated with the Lean4 ecosystem, supporting lake dependencies, Mathlib, and project-specific theorem contexts.

Can I use natural language to describe theorems?

Yes, Aristotle supports dual input, allowing you to provide English prompts within comments to generate corresponding Lean4 theorem declarations and proofs.

Aristotle Lean Theorem Prover

Name: Aristotle Lean Theorem Prover
Author: plurigrid

byplurigrid

•

Security & Testing

Automates Lean4 theorem proving and formal verification using the Harmonic API to fill proof gaps and identify counterexamples.

About

Aristotle is a high-performance Lean4 theorem prover designed for rigorous formal verification and mathematical research, achieving IMO Gold Medal level performance. By integrating with the Harmonic API, it allows developers to fill 'sorry' holes in proofs, automatically generate counterexamples for false statements, and leverage comprehensive project dependencies like Mathlib. Whether you provide natural language descriptions or existing code, Aristotle streamlines the verification process, making it an essential tool for high-assurance software development and advanced mathematics within the Claude Code environment.

Key Features

Automated hole filling for incomplete Lean4 proofs via the 'sorry' tactic
Automatic counterexample generation to debug and invalidate false statements
Seamless integration with Mathlib, lake dependencies, and project-specific theorems
Dual-input support for both English descriptions and raw Lean4 code
2 GitHub stars
Benchmarked performance with 90% MiniF2F and 96.8% VERINA scores

Use Cases

Debugging incorrect logic by generating specific counterexamples for false claims
Accelerating high-assurance software development through automated formal verification
Verifying complex mathematical conjectures and formal proofs within Lean4 environments

About

Key Features

Automated hole filling for incomplete Lean4 proofs via the 'sorry' tactic
Automatic counterexample generation to debug and invalidate false statements
Seamless integration with Mathlib, lake dependencies, and project-specific theorems
Dual-input support for both English descriptions and raw Lean4 code
2 GitHub stars
Benchmarked performance with 90% MiniF2F and 96.8% VERINA scores

Use Cases

Debugging incorrect logic by generating specific counterexamples for false claims
Accelerating high-assurance software development through automated formal verification
Verifying complex mathematical conjectures and formal proofs within Lean4 environments