# LeanProbe

> LeanProbe — epfl-lara-leanprobe. Use this tool when you need to accelerate Lean 4 proof development with fast feedback for coding agents. It solves problems of slow proof verification and environment setup by providing warm LeanInteract sessions and cached environment reuse. Ideal for use cases involving continuous integration and automated proof verification, LeanProbe accepts Lean 4 code as input and outputs verification results via its CLI, Python library, or MCP server interface.

Canonical page: https://skillsregistry.net/skills/epfl-lara-leanprobe  
JSON: https://api.skillsregistry.net/v1/skills/epfl-lara-leanprobe

## Description

Fast Lean 4 proof feedback for coding agents. CLI, Python library, and MCP server with warm LeanInteract sessions and cached env reuse.

## Trust

- **Trust score (0–1):** 0.50
- **Verification tier:** unverified

## Facts

- **Version:** 1.0.0
- **Skill type:** atomic
- **Execution layer:** container
- **Runtime environment:** vm
- **License:** MIT
- **Updated:** 2026-09-21

## Source

- **Source listing:** [GitHub](https://github.com/epfl-lara/LeanProbe)

## Use it

Resolve this record through the SkillsRegistry MCP server (no auth, read-only):

```
claude mcp add --transport http --scope user skillsregistry https://api.skillsregistry.net/mcp
```

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_skill",
    "arguments": {
      "slug": "epfl-lara-leanprobe"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/epfl-lara-leanprobe` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/epfl-lara-leanprobe/pull`

---
SkillsRegistry indexes agent skills from public registries and GitHub. Skills we have analysed are scanned with Circle-IR and scored on six dimensions; each listing states its scan coverage. More: https://skillsregistry.net/llms.txt