What is the 'consensus' tool in this skill?

The consensus tool queries multiple AI models simultaneously and analyzes their responses to find agreement, providing a more reliable and cross-checked answer for high-stakes decisions.

When should I use the AI Delegation skill?

Use it when a task requires capabilities outside Claude's primary scope, such as massive context windows (1M+ tokens), specialized math reasoning, or when you need to run models locally for privacy.

Can I use this skill without an internet connection?

Yes, the skill includes a Local-Only mode that routes all tasks to Ollama models installed on your machine, ensuring no cloud API calls are made.

How does this skill handle large codebases?

It identifies tasks requiring high context and automatically routes them to models like Gemini 3 Pro via the PAL chat tool, which can handle over a million tokens of source code.

What models are supported for delegation?

The skill supports a variety of cloud models including Gemini 1.5 Pro and Claude Opus, as well as local models via Ollama such as DeepSeek R1, Qwen, and Llama 3.

AI Delegation & Multi-Model Routing

Name: AI Delegation & Multi-Model Routing
Author: JacobPEvans

byJacobPEvans

•

数据科学与机器学习

Delegates specialized tasks to external AI models using PAL MCP tools to optimize for context size, reasoning, or privacy.

This skill empowers Claude to act as an orchestrator by routing specific tasks to external models like Gemini Pro, DeepSeek R1, or local Ollama instances when they are better suited for the job than the native model. Whether you need to process massive context windows exceeding 1 million tokens, achieve multi-model consensus on critical code reviews, or maintain strict privacy with offline-only models, this skill leverages the PAL MCP toolkit to ensure every task is handled by the most capable engine available. It streamlines complex workflows by automating the selection of specialized models for architecture planning, research, and high-stakes logic reasoning.

主要功能

01Local-only mode for private, offline development via Ollama integration

022 GitHub stars

03Multi-model consensus tools for cross-verifying critical code decisions

04Specialized task routing to models like Gemini 1.5 Pro and DeepSeek R1

05Massive context handling for tasks requiring over 1 million tokens

06Structured architecture planning using specialized 'planner' and 'clink' tools

使用场景

01Delegating complex mathematical reasoning to DeepSeek R1 for higher accuracy

02Performing full-repository architectural analysis using high-context Gemini models

03Executing sensitive code reviews locally to maintain data sovereignty and privacy

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jacobpevans/claude-code-plugins delegate-to-ai

For use in Claude.ai and ChatGPT

主要功能

01Local-only mode for private, offline development via Ollama integration

022 GitHub stars

03Multi-model consensus tools for cross-verifying critical code decisions

04Specialized task routing to models like Gemini 1.5 Pro and DeepSeek R1

05Massive context handling for tasks requiring over 1 million tokens

06Structured architecture planning using specialized 'planner' and 'clink' tools

使用场景

01Delegating complex mathematical reasoning to DeepSeek R1 for higher accuracy

02Performing full-repository architectural analysis using high-context Gemini models

03Executing sensitive code reviews locally to maintain data sovereignty and privacy