reactive Python notebooks as environments for agents

• A running marimo notebook (--no-token for auto-discovery; MARIMO_TOKEN env var for servers with auth)
• bash, curl, and jq available on PATH

Works with any agent that supports the Agent Skills open standard:

npx skills add marimo-team/marimo-pair

# or upgrade an existing install
npx skills upgrade marimo-team/marimo-pair

If you don't have npx installed but have uv:

uvx deno -A npm:skills add marimo-team/marimo-pair

Add the marketplace and install the plugin:

/plugin marketplace add marimo-team/marimo-pair
/plugin install marimo-pair@marimo-team-marimo-pair

To opt in to auto-updates (recommended), so you always get the latest version:

/plugin → Marketplaces → marimo-team-marimo-pair → Enable auto-update

The skill declares its own allowed-tools, but Claude Code may still prompt you to approve each Bash call. To avoid repeated prompts, copy the absolute paths to the scripts from the installed skill and add them to your .claude/settings.json (project-level) or ~/.claude/settings.json (global):

{
  "permissions": {
    "allow": [
      "Bash(bash /path/to/skills/marimo-pair/scripts/discover-servers.sh *)",
      "Bash(bash /path/to/skills/marimo-pair/scripts/execute-code.sh *)"
    ]
  }
}