ramazanpolat

Claude Playbooks — Development skill for Claude Code

Development community

Isolated Claude Code instances to keep things separated and clean.

How to install Claude Playbooks

This entry records only its repository, not the path inside it, so there is no exact command to give. Open ramazanpolat/claude-playbooks and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Claude Playbooks does

Isolated Claude Code instances to keep things separated and clean.

Alternatives in Development

  • Autoresearch — Claude Autoresearch Skill — Autonomous goal-directed iteration for Claude Code 2.5k ★
  • E2e Devcontainer — Run E2E tests in a dev container (isolated environment) 1.6k ★
  • Reset To Ideation — Roll back plugin to ideation stage - keep idea/mockups, remove all implementation 209 ★

README

Claude Playbooks

[![CI](https://github.com/ramazanpolat/claude-playbooks/actions/workflows/ci.yml/badge.svg)](https://github.com/ramazanpolat/claude-playbooks/actions/workflows/ci.yml)

![claude-playbook demo](docs/demo.gif)

A **Claude Playbook** is an isolated instance of Claude Code.

That's it. Each playbook has its own configuration, settings, hooks, memory, and task history -- completely separate from your default `~/.claude` installation and from every other playbook.

Why would I need this?

Claude Code stores everything in `~/.claude/`: your settings, conversation history, permissions, hooks, MCP servers. If you want to try something -- a different model, a custom hook, a new CLAUDE.md behavior -- you have to touch your main setup. One wrong change and your daily workflow breaks.

Playbooks solve this by giving each experiment (or workflow) its own isolated directory.

Common use cases:

  • Test a new hook or setting without risking your main ~/.claude
  • Separate work and personal configurations that don't interfere
  • Run two Claude Code instances concurrently on different tasks with different personalities
  • Authenticate with different accounts concurrently (e.g., keep one playbook authenticated with your corporate account and another with your personal account)
  • Share a configuration with your team by putting the playbook in a Git repo
  • Consume a repository containing one or more playbook configurations (e.g. via subdirectories)

How isolation works

Claude Code reads its configuration from the directory set in `CLAUDE_CONFIG_DIR` (defaults to `~/.claude`). Change that variable, and you get a completely fresh, independent instance:

# Your normal Claude Code (uses ~/.claude)
claude

# An isolated playbook (uses ~/.claude-playbooks/experiment)
CLAUDE_CONFIG_DIR=~/.claude-playbooks/experiment claude

That's all a playbook is under the hood. `claude-playbook` just makes creating, sharing, and managing them easy