coqtail-mcp

Drive live Rocq/Coq sessions via MCP server tools.

10|2|Updated Apr 24, 2026
One-click install
npx skills add https://github.com/park-sunho/Coqtail-mcp --skill coqtail-mcp
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: coqtail-mcp
Source: https://github.com/park-sunho/Coqtail-mcp/tree/main/skills/coqtail-mcp
Command: npx skills add https://github.com/park-sunho/Coqtail-mcp --skill coqtail-mcp

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This skill enables automated, session-based interaction with Coq via the coqtail-mcp MCP server, letting an agent drive a live Rocq/Coq proof session from file or content.

Core Features & Use Cases

  • Start and manage a Rocq/Coq session with rocq_start, reload and synchronize state with rocq_step_to, and inspect the current goal and hypotheses with rocq_goals.
  • Perform non-state-changing queries using rocq_query, or check status with rocq_status and list sessions with rocq_list.
  • Keep sessions isolated per file and control lifecycle, including closing sessions with rocq_close to avoid orphaned processes.

Quick Start

Start a session with rocq_start and then drive progression with rocq_step_to while inspecting goals via rocq_goals and issuing read-only queries with rocq_query.

Frequently Asked Questions about coqtail-mcp

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

FAQPage Schema
How do I automate a live Coq proof session through an MCP server?

Automating a live Coq proof session through an MCP server involves using dedicated tools to start sessions, step through proofs, inspect goals, and run queries. This skill drives interactive Rocq/Coq exploration by managing session lifecycle, position semantics, and error envelopes using seven MCP tools.

What's the best way to inspect Coq goals and hypotheses during interactive proof exploration?

Inspecting Coq goals and hypotheses during interactive proof exploration is done by stepping to a specific position and calling the goals tool. This skill applies read-only queries to check the current state and inspect goals without advancing the proof.

How do I manage the lifecycle of a Rocq session to avoid orphaned processes?

Managing the lifecycle of a Rocq session to avoid orphaned processes requires explicitly closing sessions when finished. This skill isolates sessions per file and controls their lifecycle from start to close using the MCP server.

Can I run read-only queries against a Coq session without changing its state?

Running read-only queries against a Coq session without changing its state is supported through the query tool. This skill performs non-state-changing operations to inspect the environment while maintaining the current proof position.

How does an agent navigate and rewind boundaries in a Coq proof using MCP?

Navigating and rewinding boundaries in a Coq proof using MCP is handled by synchronizing state to a specific position. This skill manages position semantics and rewind boundaries to ensure the agent correctly explores the proof tree.

Do I need coqidetop to start and manage Rocq sessions with MCP?

Starting and managing Rocq sessions with MCP relies on the underlying Coq interactive process. This skill starts sessions from file or content and drives the live Rocq/Coq proof session through the coqtail-mcp server.