VexJoy Agent
A pack of agents, skills and hooks with a router: the /do command picks a specialist and drives the work through checks
High risk
We rate an entry high when the tool writes to external systems, handles money, production databases or secrets, or runs arbitrary commands. The CLI installs it only with your consent.
Why this level
- Installs and runs hooks and scripts in your environment
- Runs tests, edits code and opens PRs
- The /d router calls an external service with keys
Install
Manual install
git clone https://github.com/notque/vexjoy-agent.git ~/vexjoy-agent
cd ~/vexjoy-agent
./install.shinstall.sh installs into ~/.claude/ and mirrors into ~/.codex/, ~/.factory/, ~/.reasonix/. You can choose symlinks or copies.
This is third-party code. Review the repository files before installing.
What it does
VexJoy bundles specialist agents, reusable skills, check hooks and helper scripts into one set. The /do command takes a plain-English request and selects the knowledge and tools itself, then drives the task through a pipeline: route, plan, execute with an agent plus a skill, run tests, open a branch and a PR, record the routing result. Hooks enable specific checks and scripts handle repeatable work. A separate /d router built on an external service verifies that the chosen route preserves the requested outcome. The set installs into Claude Code and mirrors into Codex, Factory and Reasonix.
Who it is for. For developers who want one entry point for tasks with automatic agent and check selection.
Good fit when
- You want to describe a task in plain English and get the right specialist
- You want a pipeline with a plan, tests and PR handling
- You want a large set of agents and skills across roles
Not a fit when
- You need one small skill without the whole environment
- You are not ready to install hooks and scripts that run commands in your environment
Example request
/do find and fix the failing test in this Go moduleLimitations
Installation places agents, skills, hooks and scripts into client home directories and runs them in your environment. The base /do router works locally, but /d needs the external TypeSafe Jev service through a direct API or the Vercel gateway with the matching keys. Some features depend on access to external services and your model provider.
How to disable. Delete the installed directories from ~/.claude/ and the mirrors in ~/.codex/, ~/.factory/, ~/.reasonix/. If you used symlinks, remove the links.
Security check
- Installs and runs hooks and scripts in your environment
- Runs tests, edits code and opens PRs
- The /d router calls an external service with keys
README in short
The README describes a set where the /do command connects plain-English requests to specialist agents, skills and workflows. It shows a six-step pipeline from routing to recording the result and an example of debugging a failing test. Installation is via git clone and ./install.sh with a choice of symlinks or copies, and the set mirrors into Codex, Factory and Reasonix. A separate /d router built on TypeSafe Jev verifies intent preservation and needs an external transport and keys. MIT licensed.
FAQ
How do /do and /d differ?
/do runs locally and picks the agent and skill itself. /d additionally verifies intent preservation through the external TypeSafe Jev service and needs keys.
Where does the set install?
Into ~/.claude/ with mirrors in ~/.codex/, ~/.factory/ and ~/.reasonix/ when the matching environment exists.
Related
Open-source personal AI assistant on your own machine: answers in Telegram, Slack, Discord and WhatsApp, extended with skills and plugins
A self-improving agent from Nous Research with a TUI, messaging gateway, cron jobs and skills it writes itself
An open source coding agent for the terminal and desktop with build and plan modes
Open prompt library with a Claude Code plugin, MCP server and CLI: search, fetch and improve prompts and skills from an agent