RootCX CLI
The RootCX CLI and skill: an app written by an agent gets a database, SSO, roles, audit and deployment on one server.
- 49 ★ on the repository
- Last commit
Broad access
Why this level
- The CLI deploys apps to a server with a shared database, secrets and employee data.
- The installer runs as a script from the network and the platform stores integration keys.
Install
- Works in
Manual install
curl -fsSL https://rootcx.com/install.sh | shCLI install for macOS and Linux from the README. On Windows use `install.ps1`.
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- The CLI deploys apps to a server with a shared database, secrets and employee data.
- The installer runs as a script from the network and the platform stores integration keys.
This is third-party code. Review the repository files before installing.
What it does
RootCX is a platform for internal tools written by an agent in Claude Code, Codex or Cursor. One server (a Docker image or the cloud) provides a shared PostgreSQL database with ready CRUD APIs, OIDC sign-in (Okta, Entra ID, Google Workspace, Auth0), RBAC roles, an audit log with before and after diffs, a secrets vault, cron, queues, file storage and more than 20 integrations. The `rootcx init` command scaffolds an app and deploys it to the cloud or your own server. The `rootcx` skill teaches the agent the platform manifest, background workers and UI components and brings references on data, integrations and agents. The core also has an MCP server, but it only reads context, validates the manifest and creates records, and never accepts application code.
Who it is for. Teams building internal tools with AI that run into access control, auditing and a place to run the apps.
Good fit when
- A department report written by an agent in one evening has to be published with role-based access.
- All the company's internal apps should live on one platform with shared SSO.
- Every data change must land in a log with its author and diff.
- Integration keys belong in an encrypted vault instead of the repository.
Not a fit when
- A public site for customers is needed: the platform targets internal tools.
- A license restricting competing products is a problem: the code is source-available under FSL-1.1-ALv2.
- The app lives on other infrastructure that cannot change.
Example request
Build a vacation request app with employee and manager roles, prepare the manifest and deploy it through RootCX.Limitations
The CLI installer runs through `curl -fsSL https://rootcx.com/install.sh | sh`, so read the script before running it. The code is licensed FSL-1.1-ALv2: you may use and modify it but not release a competing product built on it, and after two years the license becomes Apache 2.0. The skill lives in a separate repository, `rootcx/skills`. The core's MCP server has four actions in its first version, creates up to 1,000 records per call and never accepts application code.
How to disable. Delete the skill from the agent's skills folder and the `rootcx` binary from `PATH`. Remove deployed apps in the platform panel and stop the core container with Docker.
FAQ
How do I install RootCX for Claude Code?
Run `curl -fsSL https://rootcx.com/install.sh | sh` on macOS or Linux, then `npx skills add rootcx/skills`. On Windows the CLI installs with `powershell -c "irm https://rootcx.com/install.ps1 | iex"`.
What does the rootcx init command do?
It scaffolds an app, asks where to deploy (cloud or your own server) and what to name it, and deploys the app to the platform right away.
Do I have to change the tool I write code in?
No. The README promises that code is written in Claude Code, Codex, Cursor or another tool as usual, and the platform is where it runs.
What license does RootCX use?
FSL-1.1-ALv2. The code is open to read, use and modify but forbids releasing a competing product built on it. After two years it converts to Apache 2.0.
What does the RootCX MCP server do?
The first version has four tools: `get_project_context`, `get_app`, `validate_manifest` and `create_records`. It does not accept application code and does not build or deploy apps.
Discussion
No questions or reviews yet.
Related
CLI toolsThe official Coralogix CLI with 19 skills: an agent reads production logs, traces and metrics and triages incidents
Containarium: isolated boxes for AI agents
Containarium
A platform on your own server that gives each agent its own Linux box with SSH and MCP tools for commands and files
Native macOS terminal with an MCP server and CLI: agents see panes, commands and output, and you grant permissions in one window
Shows what Claude Code and Codex actually send to the API: it captures local traffic and splits tokens and cost by source
A CLI and Kubernetes operator that runs Claude Code, Codex, Gemini, OpenCode and Cursor as tasks and sessions inside a cluster
A self-hosted control center for coding agents: Claude Code, Codex, Copilot, OpenCode and Pi from your phone, desktop and terminal
README in short
The README opens with a problem: an app written by an agent hits access, database and audit questions within five minutes. It then covers installing the CLI and skill, a comparison with a Supabase plus Auth0 setup and with Retool, an architecture diagram and a feature table.