hol-light

Name: hol-light
Availability: InStock
Author: jrh13

Automate HOL Light theorem proving with MCP tools for goal setting and tactic application.

547|100|Updated Oct 19, 2015

Search Tags:#automation #mcp #proof-assistant #ocaml #theorem-prover #hol-light #interactive-proof

Ready to run

npx skills add https://github.com/jrh13/hol-light --skill hol-light

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: hol-light
Source: https://github.com/jrh13/hol-light/tree/main/mcp
Command: npx skills add https://github.com/jrh13/hol-light --skill hol-light

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Automate interactive theorem proving for HOL Light via MCP tools to streamline proof construction, verification, and lemma management within neural-assisted workflows.

Core Features & Use Cases

Interactive proof workflow using set_goal, goal_state, search_theorems, apply_tactic, apply_tactics, backtrack, and hol_help.
One-shot proving via prove and prove-like wrappers, plus tools for goal inspection, theorem search, and type queries.
Proof recording, timeout control, checkpoint restart, and session management to support reproducible experiments.

Quick Start

Launch the HOL Light MCP server and connect with the client, then set a goal and begin proving.

Frequently Asked Questions about hol-light

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

FAQPage Schema

How do I automate HOL Light theorem proving in interactive sessions?▼

Automating interactive theorem proving for HOL Light involves using MCP tools to manage goal states, apply tactics, and search for lemmas. It exposes commands such as set_goal, goal_state, apply_tactic, and backtrack to streamline proof construction and verification.

What commands are available for applying tactics and searching lemmas in HOL Light?▼

The available commands include set_goal, goal_state, search_theorems, apply_tactic, apply_tactics, prove, backtrack, hol_help, hol_status, hol_load, and hol_type, enabling goal inspection, tactic application, theorem discovery, and proof reconstruction.

Can I use MCP tools to manage HOL Light proof checkpoints and timeouts?▼

Yes, MCP tools support proof recording, configurable timeouts, checkpoint restarts, and session management for HOL Light, ensuring reproducible experiments and controlled execution limits during automated proof search and verification.

Does the HOL Light MCP server support one-shot proving and proof recording?▼

Yes, the server supports one-shot proving via the prove command and proof-like wrappers, alongside proof recording features to capture and replay proof steps for reproducible neural-assisted workflows.

What is the best way to integrate HOL Light with neural-assisted proof workflows?▼

The best way to integrate HOL Light with neural-assisted workflows is by connecting to the MCP server, which exposes commands for goal setting, tactic application, lemma discovery, and proof reconstruction to streamline automated theorem verification.