pi

Route user-defined tasks and skills through Pi for comparison with Claude Code.

Updated Jun 4, 2026
One-click install
npx skills add https://github.com/hemory/amp --skill pi-hemory
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: pi
Source: https://github.com/hemory/amp/tree/main/.claude/reference/beta-templates/pi/skills/pi
Command: npx skills add https://github.com/hemory/amp --skill pi-hemory

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill allows you to directly compare the performance and output of different AI coding agents, specifically Pi and Claude Code, by routing tasks through each.

Core Features & Use Cases

  • Direct Comparison: Run the exact same task or skill through Pi and Claude Code to evaluate differences.
  • Mode Toggling: Easily switch between using Pi as the primary agent or reverting to Claude Code.
  • Use Case: You're developing a new feature and want to see which AI agent generates better code. Use /pi /build-feature-x to run it through Pi, then run /build-feature-x in Claude Code to compare the results.

Quick Start

Run the /daily-plan skill through Pi for comparison.

Frequently Asked Questions about pi

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

FAQPage Schema
How do I compare AI coding agents side-by-side for the same task?

Comparing AI coding agents side-by-side involves routing the exact same task through different tools to evaluate output differences. This Skill facilitates direct comparison between Pi and Claude Code by executing predefined skills and user-defined tasks through each agent.

Can I switch between Pi and Claude Code as my primary coding agent?

Yes, you can switch between Pi and Claude Code as your primary coding agent. The Skill provides mode toggling functionality, allowing you to easily revert to Claude Code or set Pi as the primary agent for your workflow.

Do I need Pi installed to use this AI comparison workflow?

Yes, you must have Pi installed to use this AI comparison workflow. Activating the comparison requires a local Pi installation, and you may also need to activate specific beta features to route tasks and predefined skills correctly.

What is the best way to evaluate which AI agent generates better code for a new feature?

The best way to evaluate which AI agent generates better code is to run the exact same feature task through both agents. You can run `/pi /build-feature-x` through Pi, then run `/build-feature-x` in Claude Code to directly compare the results.

How does routing predefined skills through Pi work?

Routing predefined skills through Pi works by passing specific prompts and context to the Pi agent using a command like `/pi /daily-plan`. This launches Pi with the exact instructions you define, enabling a direct performance comparison against Claude Code.