consult-llm

Coordinate external LLM consultations by gathering context and invoking the consult_llm MCP tool.

126|22|Updated Jun 22, 2025
One-click install
npx skills add https://github.com/raine/consult-llm-mcp --skill consult-llm
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: consult-llm
Source: https://github.com/raine/consult-llm-mcp/tree/main/examples
Command: npx skills add https://github.com/raine/consult-llm-mcp --skill consult-llm

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill automates the process of consulting external LLMs by collecting relevant file context, selecting the appropriate model/mode (Gemini or Codex, API/CLI/web), and invoking the consult_llm MCP tool. It ensures neutral prompts and focused context to reduce noise and save you time.

Core Features & Use Cases

  • Gather and filter relevant files automatically to provide concise context for LLMs.
  • Decide the mode and model based on user instruction ("ask gemini" uses Gemini; "ask codex" uses Codex; "ask in browser" uses web mode).
  • Call the MCP tool consult_llm with API, CLI, or Web modes and present results.
  • Centralized logging and consistent output formatting to simplify comparisons and decisions.

Quick Start

In Claude Code, say: "ask gemini about the race condition in server.ts" or "ask codex about the race condition in server.ts" or "ask in browser about the race condition in server.ts". The skill will gather relevant files, determine mode and model, and present a summarized response with next steps.

Frequently Asked Questions about consult-llm

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

FAQPage Schema
How do I consult external LLMs with relevant file context automatically?

Consult external LLMs by gathering repository context, selecting the appropriate model (Gemini or Codex), and invoking the consult_llm MCP tool with filtered files. This automates context collection and reduces noise in your query.

Can I ask Gemini or Codex about specific code problems in my repository?

Yes. Use commands like 'ask gemini about the race condition in server.ts' or 'ask codex about the issue in main.py'. The Skill gathers relevant files, determines the model, and presents a summarized response with next steps.

What's the difference between API, CLI, and web modes for LLM consultation?

API mode invokes Gemini or Codex programmatically; CLI mode runs through command-line; web mode consults in the browser. The Skill automatically selects the mode based on your instruction ('ask in browser' triggers web mode) and your environment.

How does the Skill filter files to provide focused context?

The Skill uses Glob and Grep to automatically gather and filter relevant files from your repository, then prioritizes file relevance to create concise context. This reduces noise and ensures the LLM receives only pertinent information.

Do I need to manually format prompts before consulting an LLM?

No. The Skill enforces neutral, focused prompts and standardized output formatting automatically. It handles context collection and prompt structuring so you can focus on the core question.