pair-with-me

Resolve GitHub issues collaboratively with AI and human oversight.

Updated Jun 22, 2026
One-click install
npx skills add https://github.com/bastien-gallay/pair-with-me --skill pair-with-me
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: pair-with-me
Source: https://github.com/bastien-gallay/pair-with-me/tree/main
Command: npx skills add https://github.com/bastien-gallay/pair-with-me --skill pair-with-me

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires GitHub API, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill fosters a collaborative workflow between humans and AI for GitHub issues, reducing cognitive debt and ensuring human authorship remains intact.

Core Features & Use Cases

  • Adversarial Dialogue: Encourages a structured back-and-forth where the human produces content before the AI reveals its insights.
  • Test-First Approach: Implements a Red→Green→Refactor→Reflect loop for each test, emphasizing quality over quantity.
  • Two Pauses: Ensures human involvement in key decision points, maintaining oversight and accountability.
  • Async Gates: Interrupts where human judgment is crucial, maintaining focus and efficiency.
  • Retention Pass: Provides a visual recap and novelty ledger to reinforce learning and improve retention.

Quick Start

Use the /pair-with-me command followed by the GitHub issue number, e.g., /pair-with-me #123.

Frequently Asked Questions about pair-with-me

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

FAQPage Schema
How do I use AI to collaboratively resolve GitHub issues while maintaining human authorship?

You can collaboratively resolve GitHub issues using AI by triggering a workflow with an issue number, which employs adversarial dialogue and async gates to ensure human oversight. The AI interacts with the GitHub API to handle issue metadata while you retain authorship.

What is the best way to reduce cognitive load when tackling technical debt on GitHub?

To reduce cognitive load when tackling technical debt, you can use a collaborative AI workflow that shares issue resolution tasks. It employs a test-first approach and structured feedback loops, interrupting at key decision points to maintain your focus and accountability.

Can I integrate test-first development with AI assistance for GitHub issue resolution?

Yes, you can integrate test-first development with AI assistance by running a Red, Green, Refactor, and Reflect loop for each test. This approach emphasizes quality over quantity and ensures structured human-AI collaboration throughout the issue resolution process.

Do I need GitHub API access to run adversarial dialogue with an AI for issue resolution?

Yes, you need GitHub API access to run adversarial dialogue with an AI for issue resolution. The AI requires API interaction to fetch and handle issue metadata, enabling the structured back-and-forth where you produce content before the AI reveals its insights.

How does the structured feedback loop work during AI-assisted technical debt reduction?

The structured feedback loop works by implementing two pauses, or async gates, where human judgment is crucial. It also includes a retention pass with a visual recap and novelty ledger to reinforce learning, ensuring human involvement in key decision points.

Why does AI collaboration interrupt my workflow during GitHub issue resolution?

AI collaboration interrupts your workflow during GitHub issue resolution to enforce async gates, ensuring human judgment is applied at critical decision points. These pauses maintain oversight and accountability, preventing the AI from autonomously resolving issues without human authorship.