goldenfox2025

Fenn Code — AI skill for Claude Code

AI community

A desktop workspace for the AI agents you already use: chats, projects, threads, scheduled tasks and computer control on top of ACP agents (Codex, Claude Code, Grok, Antigravity, Qoder).

How to install Fenn Code

This entry records only its repository, not the path inside it, so there is no exact command to give. Open goldenfox2025/fenn-code and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Fenn Code does

A desktop workspace for the AI agents you already use: chats, projects, threads, scheduled tasks and computer control on top of ACP agents (Codex, Claude Code, Grok, Antigravity, Qoder). Unofficial; not affiliated with Anthropic.

Alternatives in AI

Open Dots — Open-source, self-hosted AI agent workspace and alternative to OpenAI Dots, Meta Muse, Grok Bot, Instinct, Man 5.5k ★
Azure AI Agents .net — Persistent AI agents with threads and tools 1.8k ★
Gh PR Review — GitHub CLI extension that adds full inline PR review comment support — view, navigate, reply to, and resolve r 174 ★

README

Fenn Code

**A desktop workspace for the AI agents you already use.** Chats, project threads, interactive replies, scheduled tasks and computer control, with Codex, Claude Code, Grok, Qoder, Antigravity or another agent that speaks the [Agent Client Protocol](https://agentclientprotocol.com) (ACP).

[简体中文](README.zh.md) · [Download](https://github.com/goldenfox2025/fenn-code/releases) · [Build from source](#build-from-source)

![Fenn Code: a project chat with its threads, and one thread asking which headline to use](docs/screenshots/en/hero.webp)

Why Fenn Code

Command-line agents are good at the work but leave you with a terminal per task. Fenn Code gives them the structure of a desktop app: one place to ask, threads that keep long work apart, a record of what was decided and what was produced, and tasks that run on a schedule. The agent stays your choice, and switching agents doesn't lose the project.

Features

Projects with threads

A project has its own chat, run by a coordinator agent. It answers small questions itself and starts a **thread** for anything that takes real work. Each thread keeps a live checklist, names its current obstacle, and asks for a decision with a card when it needs one. An overview shows what is waiting on you, what is in progress and what is done, next to a progress board of milestones the agents keep up to date.

Research and investigation threads share the project's context and folder. Code-changing threads receive their own branch and worktree from the project's committed HEAD; an investigation can request a code workspace and continue in the same thread. Uncommitted project changes are excluded from new worktrees. A worker submits its delivery and automatically triggers coordinator review of the actual changes, evidence, overlaps and integration impact. A compact review card sits below the thread's history and above its composer. Open its details for analysis, the integration plan and changed files, then **Approve and