lean-collab

Orchestrate parallel Lean proof sessions via the lc CLI.

72|2|Updated Jan 10, 2026
One-click install
npx skills add https://github.com/mutable-state-inc/lean-collab --skill lean-collab
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: lean-collab
Source: https://github.com/mutable-state-inc/lean-collab/tree/main/skills/lean-collab
Command: npx skills add https://github.com/mutable-state-inc/lean-collab --skill lean-collab

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Collaborative theorem proving often stalls as humans juggle goals, decompositions, and verification steps. Lean-collab orchestrates multi-agent proof sessions, managing skeleton design, verification, and final proof composition without manual micromanagement.

Core Features & Use Cases

  • Orchestrates parallel Lean proof sessions via a CLI workflow (lc) to run decomposers and provers concurrently.
  • Maintains a structured workflow from root theorem initialization through status checks, decompositions, backtracking, verification, and final composition.
  • Supports real-time, Ensue-coordinated status updates and goal management for collaborative formalization tasks.

Quick Start

Initialize the LC session with a root theorem and hypotheses, then let the orchestrator run until the proof is complete.

Frequently Asked Questions about lean-collab

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I coordinate parallel Lean proofs across multiple agents?

Parallel Lean proofs are coordinated by orchestrating multi-agent sessions via the lc CLI, which runs decomposers and provers concurrently while tracking real-time status to produce a verified proof.

What is multi-agent theorem proving and when do I need it for Lean projects?

Multi-agent theorem proving decomposes a root theorem into parallel sub-goals handled by concurrent provers, needed when formalization tasks outpace single-agent verification or require structured skeleton design and backtracking.

How do I start a collaborative Lean proof session using a CLI workflow?

Initialize the session by declaring a root theorem and hypotheses through the lc CLI, then let the orchestrator manage decomposition, verification, and final proof composition until completion.

Can I manage Lean proof goals and backtracking automatically during multi-agent sessions?

Automated goal management and backtracking are supported natively, allowing the orchestrator to track sub-goal status, apply decompositions, and revert failed proof branches without manual intervention.

Does lean-collab work without external dependencies for theorem verification?

Lean-collab operates with zero external dependencies, relying solely on the lc CLI and Ensue-coordinated execution to manage the full proof lifecycle from initialization through final composition.

Why does my collaborative Lean proof workflow stall during goal decomposition?

Proof workflows stall when decomposition, verification, and composition steps lack coordinated status tracking; lean-collab resolves this by orchestrating parallel agent tasks with real-time Ensue updates and automated goal management.