CLAUDE LABJP
FORK — Claude Code 2.1.212 changes what /fork does: it copies your conversation into a new background session with its own row in claude agents, so you can keep working. The old in-session subagent is now /subtaskLIMITS — WebSearch calls are now capped at 200 per session by default, and subagent spawns get the same 200 ceiling, so a runaway search or delegation loop stops on its ownMCPBG — MCP tool calls running past two minutes now move to the background automatically, keeping the session usable. Tune the threshold with CLAUDE_CODE_MCP_AUTO_BACKGROUND_MSPLANFIX — Fixed plan mode auto-running file-modifying Bash commands such as touch and rm without a permission prompt or an SDK canUseTool callbackSONNET5 — Claude Sonnet 5 is running on introductory pricing of $2 per million input tokens and $10 per million output. After August 31 it moves to $3 and $15IPO — Bankers are reportedly lining up investor meetings for Anthropic ahead of a possible public listing as soon as OctoberFORK — Claude Code 2.1.212 changes what /fork does: it copies your conversation into a new background session with its own row in claude agents, so you can keep working. The old in-session subagent is now /subtaskLIMITS — WebSearch calls are now capped at 200 per session by default, and subagent spawns get the same 200 ceiling, so a runaway search or delegation loop stops on its ownMCPBG — MCP tool calls running past two minutes now move to the background automatically, keeping the session usable. Tune the threshold with CLAUDE_CODE_MCP_AUTO_BACKGROUND_MSPLANFIX — Fixed plan mode auto-running file-modifying Bash commands such as touch and rm without a permission prompt or an SDK canUseTool callbackSONNET5 — Claude Sonnet 5 is running on introductory pricing of $2 per million input tokens and $10 per million output. After August 31 it moves to $3 and $15IPO — Bankers are reportedly lining up investor meetings for Anthropic ahead of a possible public listing as soon as October
Articles/Claude Code
Claude Code/2026-05-06Intermediate

How gh skill Lets You Share SKILL.md Across Claude Code, Copilot, Cursor, and 30+ AI Agents

The gh skill GitHub CLI extension lets you publish and install SKILL.md definitions across Claude Code, GitHub Copilot, Cursor, Gemini CLI, and 30+ AI agents — write once, deploy everywhere.

Claude Code196SKILL.md6gh skill2GitHub CLIAI agents2CopilotCursor4

If you work on a team where different developers use different AI coding agents, you've probably hit this wall: the SKILL.md you carefully crafted for Claude Code does nothing for your teammate who lives in GitHub Copilot. The knowledge stays siloed.

That's exactly the problem the gh skill GitHub CLI extension is designed to solve. It lets you package SKILL.md-based skills and distribute them across multiple AI agents — so the standards and patterns you define actually reach the whole team, regardless of which tool each person prefers.

The Problem With Agent-Specific Skill Files

Claude Code's SKILL.md is genuinely useful. You can encode code review workflows, project-specific conventions, and implementation patterns that Claude will consistently follow. But it's a Claude Code artifact. GitHub Copilot, Cursor, and Gemini CLI each have their own configuration conventions.

The result: teams end up maintaining parallel files for the same intent. One person keeps .cursorrules up to date, another writes Copilot instruction snippets, and the Claude Code users have SKILL.md. All describing the same thing, drifting apart slowly.

What gh skill Does

gh skill is a GitHub CLI extension that turns SKILL.md files into distributable packages. A skill package lives in a public GitHub repository, and any developer can install it with a single command. The extension handles translating the skill definitions into whatever format each target agent expects.

Supported agents as of 2026 include Claude Code, GitHub Copilot, Cursor, Codex CLI, Gemini CLI, and 30+ community-maintained agents like Aider, Continue, and Cline.

Getting Started

You'll need the GitHub CLI installed. Then:

# Install the gh skill extension
gh extension install github-actions/gh-skills
 
# Verify installation
gh skill --version

To install a skill from a public repository:

# Install for all supported agents
gh skill install example-org/coding-skills
 
# Install for a specific agent only
gh skill install example-org/coding-skills --agent claude-code
gh skill install example-org/coding-skills --agent copilot

Structuring Your SKILL.md for gh skill

Your existing SKILL.md content can mostly stay as-is. What you add is a skill.yaml metadata file that describes which agents receive which skill files, and optionally a version tag.

# skill.yaml
name: my-project-skills
version: 1.0.0
description: "Coding standards and workflows for the XYZ project"
agents:
  - claude-code
  - copilot
  - cursor
skills:
  - path: skills/code-review.md
    agents: [claude-code, copilot, cursor]
  - path: skills/deployment.md
    agents: [claude-code]  # Deployment steps only for Claude Code

The referenced Markdown files follow the same SKILL.md conventions you already know. Each agent interprets the content according to its own capabilities — Claude Code handles the richest subset, while agents like Copilot parse more general instructions.

Publishing a Skill for Your Team

To share your skills with teammates or the broader community, organize your repository like this:

my-skills-repo/
├── skill.yaml
├── README.md
└── skills/
    ├── code-review.md
    ├── testing.md
    └── deployment.md

Push to a public GitHub repository, and your team can install with:

gh skill install your-username/my-skills-repo

To stay in sync, teammates run:

gh skill update my-project-skills

For version pinning — useful when you want to ensure a stable baseline before upgrading:

gh skill install your-username/my-skills-repo@v1.0.0

What Actually Changes in Practice

The main shift is that "update the project standards" becomes one operation instead of four. When you refine how you want code reviews to work, you push to the skill repo, cut a new version, and teammates sync up. The intent stays unified across agents.

One practical caveat: interpretation fidelity varies by agent. Claude Code handles the most nuanced instructions — tool restrictions, hook definitions, multi-step workflows. GitHub Copilot reads the Markdown but applies it more loosely. When writing for cross-agent distribution, keep critical instructions in plain, direct language so they survive the translation to simpler agents.

Start With One Skill

You don't need a fully structured skill library to get started. Pick one workflow that matters to your team — code review guidelines, commit message conventions, testing patterns — and package it. Once the mechanics click, the rest follows naturally.

gh extension install github-actions/gh-skills
gh skill list

For deeper implementation patterns — versioning strategies, CI/CD integration, and agent-specific optimization — the premium article covers the full architecture.

Share

Thank You for Reading

Claude Lab is ad-free, supported entirely by members like you. We publish practical guides daily with implementation code, benchmarks, and production-ready patterns. If you've found it useful, we'd love to have you on board.

  • Copy-paste ready implementation code
  • New advanced guides published daily
  • $5/mo or $10 for lifetime access
View Membership →

If you found this article helpful, a small tip ($1.50) would mean a lot to us. Your support helps keep this site ad-free and covers server and hosting costs.

Related Articles

Claude Code2026-05-06
Multi-Agent Skill Architecture with gh skill — Versioning, CI/CD, and Agent-Specific Optimization
A production-grade approach to managing SKILL.md across teams using gh skill. Covers repository structure, semantic versioning, GitHub Actions automation, agent-specific optimization, and monorepo patterns.
Claude Code2026-07-09
Claude Code vs Cursor: The Definitive 2026 Comparison — Choosing the Right AI Coding Tool
A comprehensive comparison of Claude Code and Cursor across pricing, features, accuracy, and workflow. Find the AI coding tool that best fits your development style with our 2026 data-driven guide.
Claude Code2026-06-13
When Your Edited SKILL.md Doesn't Take Effect — Hot-Swapping Claude Code Skills with /reload-skills and Auto-Loaded .claude/skills
A practical routine for hot-swapping Claude Code skills without restarts: /reload-skills, SessionStart hooks, and version stamps that show which SKILL.md is live.
📚RECOMMENDED BOOKS
Build a Large Language Model (From Scratch)
Sebastian Raschka
LLM Dev
Prompt Engineering for LLMs
Berryman & Ziegler
Prompting
AI Engineering
Chip Huyen
AI Eng
* Contains affiliate links
See all →