Agent Skills that teach AI coding assistants (Claude Code, Cursor, Codex) how to deploy, operate, and debug applications on BrainPod using the brainpod CLI.
skills/brainpod covers the whole lifecycle: creating pods, building and pushing container images, composing resources, promoting draft revisions, changing what a running pod runs, and diagnosing unhealthy deployments.
It routes to use-case references it loads on demand:
references/deploy.md— take an application from source to a running pod for the first time.references/operate.md— change what an existing pod runs: resources, env, scale, databases, blueprints, and reverting a bad deploy.references/debug.md— diagnose a deployment that failed, timed out, is unhealthy, or is not serving traffic.
Open your coding agent in the project you want to ship and paste:
Install the BrainPod skill from github.com/brainpodnl/skills and use it to deploy this project to BrainPod.
Sign me up, work out what the project needs, and hand me the live URL when it's up.
Ask your agent to install it, or run one of the following yourself.
Install as a Cursor plugin:
/add-plugin brainpodInstall via the skills CLI:
npx skills add brainpodnl/skills --skill "brainpod"Clone this repo and symlink the skill into your agent's skills directory:
git clone https://github.com/brainpodnl/skills.git /path/to/brainpod-skillsln -s /path/to/brainpod-skills/skills/brainpod /path/to/skills-directory/brainpodYou don't need to set anything up before you start. Ask the agent to deploy something to BrainPod and it handles onboarding for you — installing the CLI if it is missing, and signing you in.
For reference, what it sets up on your behalf:
The brainpod CLI, from brainpodnl/cli. Linux and macOS on amd64 and arm64.
A signed-in session. The agent runs brainpod login and opens the sign-in page for you — in an embedded browser pane where its harness has one, otherwise in your default browser. You approve there and the CLI stores the token itself, so there is no token to copy and paste. Starting the flow is the agent's part; approving it is yours.
On a headless machine, in a container, or in CI, the browser flow isn't available, so create an API token in the BrainPod console and set it instead:
export BRAINPOD_API_TOKEN=brain_...The CLI also accepts --api-token and brainpod config set api-token <token>. Configuration lives at ~/.config/brainpod/config.toml.
Docker with Buildx, required only for image builds.
Once installed, the agent uses the skill automatically when relevant — for example:
- Deploying a repository to BrainPod from scratch
- Adding a Postgres, MariaDB, or Valkey database to an existing pod
- Changing environment variables, replica counts, or instance sizes
- Installing a blueprint and pinning it to a manually built image
- Reverting a bad deploy
- Diagnosing a deployment that reports as deployed but is not responding
Authoring and review guidance lives in AGENTS.md. It is the source of truth for how skill content in this repo is written and reviewed.