myrlin-workbook

Name: myrlin-workbook
Author: therealarthur

Verified

Open-source workspace manager for AI coding CLIs: Claude Code and ChatGPT Codex in one browser app. Discovers every session on disk, embedded terminals, phone-ready, cost tracking, kanban, docs. Windows, macOS, Linux.

371stars

51forks

JavaScript

Installation

# Add to your Claude Code skills
git clone https://github.com/therealarthur/myrlin-workbook

Getting Started

Guides for using ai agents skills like myrlin-workbook.

Caveman: Cut Claude Token Use by 65%
How agent-side prompt compression works, when to use it, and when not to.
What is an AI Skills Marketplace?
Definitions, how marketplaces work, and how to choose between them in 2026.
Getting Started with AI Skills

Security ReportVerified

Last scanned: 8/19/2026

{
  "issues": [],
  "status": "PASSED",
  "scannedAt": "2026-08-19T04:35:50.418Z",
  "npmAuditRan": true,
  "pipAuditRan": true,
  "promptInjectionRan": true
}

README.md

Frequently Asked Questions

What is myrlin-workbook?

myrlin-workbook is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by therealarthur. Open-source workspace manager for AI coding CLIs: Claude Code and ChatGPT Codex in one browser app. Discovers every session on disk, embedded terminals, phone-ready, cost tracking, kanban, docs. Windows, macOS, Linux. It has 371 GitHub stars.

Is myrlin-workbook safe to use?

Yes. myrlin-workbook passed SkillsLLM's automated security scan — a dependency vulnerability audit plus prompt-injection heuristics — with no high-severity issues. You can read the full report in the Security Report section on this page.

How do I install myrlin-workbook?

Clone the repository with "git clone https://github.com/therealarthur/myrlin-workbook" and add it to your Claude Code skills directory (see the Installation section above).

What programming language is myrlin-workbook written in?

myrlin-workbook is primarily written in JavaScript. It is open-source under therealarthur on GitHub, so you can review or fork the full source.

Are there alternatives to myrlin-workbook?

Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh myrlin-workbook against similar tools.

Agentic AI for Beginners

Build your first AI agent from scratch - tool use, ReAct pattern, memory, deployment

41 minBeginner

Comments (0)

to leave a comment.

No comments yet. Be the first to share your thoughts!

Related Skills

ECC

by affaan-m

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

241,003

mobileClaw tfmcp

⚠️ Third-Party Software Notice

This skill is third-party open-source software developed and hosted independently on GitHub. SkillsLLM is an informational directory and does not control or maintain the underlying repository.

Any security checks, ratings, or warnings displayed by SkillsLLM are automated and limited in scope. They do not constitute a security certification or guarantee that the software is safe, error-free, or free from malicious code, vulnerabilities, compromised dependencies, or prompt-injection risks.

Review the source code, permissions, dependencies, and configuration before installing or running any third-party skill. Use is at your own risk. To the maximum extent permitted by applicable law, SkillsLLM is not liable for losses arising from third-party software.

Read the Terms of Service

Myrlin Workbook

Every Claude Code and ChatGPT Codex session on your machine, in one sidebar.

Install · Features · Phone · Codex · Architecture · Docs · Contributing

Why

I run Claude Code all day and I started running Codex alongside it. Between the two there are a few hundred conversations on this machine, spread across two CLIs that do not know about each other, in a dozen project folders, and the only way to get back to one is to pick a UUID out of a list. Open three at once and you are juggling terminal windows. Restart the machine and you reopen everything by hand. Scroll up in a pane to find what an agent did an hour ago and you hit the top of the buffer, because a coding agent repaints its screen and keeps no scrollback at all.

So I built the thing I wanted. Myrlin reads both CLIs where they already store their data, groups every session by the folder it ran in, and opens any of them in a real terminal in a browser tab. You can wheel back through the whole conversation, drag a selection across an hour of history and copy it. Costs come out of the transcripts. Docs, a kanban board and git worktrees sit next to the sessions they belong to. It runs on your phone over your own network.

It is not a hosted service and it is not an agent. There is no sign-up, no cloud, no telemetry, and no model gets called unless you ask for one. It is a local Node server, a browser tab, and your own two CLIs doing what they already do.

Install

npx myrlin-workbook@alpha    # v1.3 alpha: Claude Code and ChatGPT Codex, redesigned, phone layout
npx myrlin-workbook          # v0.9 stable: Claude only, unchanged
npx myrlin-workbook --demo   # sample data, no real sessions needed

Node 20 or newer. node-pty needs C++ build tools; if they are missing the app still boots, just without terminal panes.

A random password is generated on first launch and saved to ~/.myrlin/config.json, where it survives updates, reinstalls and npx cache clears. Set CWM_PASSWORD to override it.

The server binds to 127.0.0.1 by default. Set CWM_HOST to your LAN IP to reach it from a phone or another machine, and set a real password first.

Everything else, including tunnels, environment variables, keyboard shortcuts and troubleshooting, is in SETUP.md.

Features

Every session, already found

Claude Code sessions are read from ~/.claude/projects. Codex sessions are read from ~/.codex, including state_5.sqlite, which is the same thread store the ChatGPT desktop app lists from, so the folders and sessions you see here are the ones that app shows you. Both reads are read-only. Sessions group by the directory they ran in, with a Recent list at the top and a Discovered section at the bottom for anything on disk you have not filed yet. The All / Claude Code / ChatGPT Codex switcher sits inside that section and filters that list only. Ctrl+K finds anything by name.

A terminal you can scroll back through

xterm.js over node-pty, so it is a real PTY. On Windows that is ConPTY directly, with no WSL; macOS and Linux use their own. Wheel up past the top of what a pane has drawn and it carries on into the recorded conversation: same background, same typeface, no panel to open and nothing to switch on. Drag from the line being written right now up into last hour's output and it is one continuous selection, which Ctrl+C copies. Ctrl+Shift+A takes the whole document, Shift+PageUp pages, Escape leaves, and typing anything leaves and types. Output keeps flowing while you read.

ChatGPT Codex, read from its own store

Codex threads come out of the SQLite database the desktop app uses, through an in-memory byte image, so no write handle is ever opened against your session history. A filesystem walk over the rollout JSONL stays as a fallback for machines the desktop app has never run on. The detail strip shows the model, reasoning effort, approval policy and sandbox that the session is actually running, read from the conversation itself; anything genuinely unknown says unknown rather than being filled in with a guess. Codex reports token counts and no dollar figure, because there is no published price to apply and inventing one would be worse than showing nothing.

It works on a phone

Five tabs: Home, Sessions, Terminal, Attention and Search. Attention carries the only persistent badge in the app, counting what is waiting on you. Flick up past the top of a terminal and the history opens and keeps going; a long press hands you the phone's own selection handles and Copy bar rather than an imitation of them. The key row fits five keys plus an overflow menu on a 390 pixel screen, and holding Ctrl+C gives you every control key from Ctrl+A to Ctrl+Z. The layout is driven by visualViewport, so the terminal stays above the soft keyboard. Reach it over your LAN with CWM_HOST, or through a Cloudflare tunnel the app can start for you.

Thirteen terminal palettes, light and dark chrome

Two independent choices. The app chrome is light or dark, toggled from the topbar. The terminal palette is one of thirteen: Mocha, Macchiato, Frappé, Nord, Dracula, Tokyo Night, Cherry, Ocean, Amber and Mint in the dark set, Latte, Rose Pine Dawn and Gruvbox Light in the light one. Pick it under Settings, Interface, Terminal theme, with a live preview swatch; the choice persists in localStorage. Every palette is derived from the same CSS custom properties as the chrome, and a test fails if the two ever drift apart.

What you spent, and what is next

Cost comes out of Claude's own usage fields in the transcripts: input, output, cache write and cache read tokens, priced per model, per session and per project. Parsing runs on a worker thread so a large transcript never stalls terminal I/O. The Costs view has a Day, Week, Month and All selector, a timeline chart and a table you can sort by cost, tokens or duration. The agent task board runs Backlog, Planning, Running, Review and Done with drag and drop, and a card can own a git branch, a worktree and a live session, then open a pull request through gh and track its state on the card.

Session templates. Save a working directory, model, flags and spawn options, then launch from it in one click.
**Conflict