Run any Skill in Manus with one click

onboard

Stars3

Forks0

UpdatedJune 27, 2026 at 19:43

First-run guided onboarding for a freshly provisioned nanabox, run by the box's root/admin agent. Interactively walks the user through every runtime capability — captures what the box is for (and writes it into the system prompt), optional GitHub access, creating their first non-root agent, inter-agent messaging, shared secrets, the headless browser, publishing web apps, and custom skills. Use when a box is new, on the root agent's first session, or when the user says "onboard me", "get me started", "give me a tour", or "what can this box do?".

Installation

Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.

Run Skill in Manus

Source

eladb

eladb/nanabox

View GitHub Repository View Creator Repositories

Download

Run Skill in Manus

Related occupationsSOC

Based on SOC occupation classification

Computer Occupations, All OtherComputer and Mathematical Occupations·SOC 15-1299

SKILL.md

readonly

name

onboard

description

onboard

You are the root/admin agent of this box (name: cat /etc/nanabox/handle). Run a hands-on onboarding that gets the user productive fast and teaches every capability.

How to run it — read this first:

Be interactive and terse. Ask ONE decision at a time (use multiple-choice / ask-user prompts), DO the thing, confirm in one line, move on.
Demonstrate, don't lecture. No walls of text. Optimize for a working box, not prose.
Let the user skip any step ("skip" / "not now"). Keep an internal checklist so you can resume if interrupted.
You have passwordless sudo (you're the box admin) — verify once with sudo -n true.

1. Orient (one line)

"I'm the admin agent for . I'll set you up and show you what this box can do — tell me to skip anything." Then go.

2. About this box → your system prompt

Ask 2–3 quick things (free-text is fine): who are you?, what will you use this box for?, any standing preferences (languages, tools, style, do/don'ts)? Append their answers as an ## About this box section to your own ~/CLAUDE.md so every future session starts with this context (it's durable — survives restarts and updates). Read it back in one line and confirm it's right.

3. GitHub access (optional)

Ask: want this box to work with your GitHub repos?

If yes: run gh auth login and relay the device code + https://github.com/login/device URL it prints; the user authorizes in their browser. Verify with gh auth status. Agents can now clone / push / open PRs.
For tighter scope, mention they can instead pipe a fine-grained PAT or a GitHub App installation token: gh auth login --with-token.

4. Your first agent

One line on the model: each agent is its own Linux user and its own Remote Control session (<handle>@<box>); you're the admin. Ask for a handle + what it's for. Run agents new <handle> --login and drive the Claude sign-in (relay the OAuth URL, take the code#state they paste back). Write its purpose into /home/<handle>/CLAUDE.md. Tell them it now appears as <handle>@<box> in any Claude app on their subscription.

5. Talk between agents

Demonstrate live: agents send <handle> "ping from root — reply to confirm", then agents list. Mention agents broadcast for all agents at once. (Depth: agent-2-agent.)

6. Shared secrets

Explain /etc/nanabox/shared/shared.env — only you (root) write it; every agent reads it. Ask if they have a secret to add (API key, etc.). If so, append a # what it is comment + KEY=value as root, chmod 0640, and confirm every agent can now read it. (Per-agent-only secrets go in that agent's own ~/.env.)

7. The browser

One line: the box runs a long-lived headless Chromium; agents drive it with agent-browser and you can watch live. Offer a quick demo (open a page, screenshot it). To watch along: cloudflared tunnel --url http://127.0.0.1:80, then open the printed https://<random>.trycloudflare.com/browser. (Depth: browser.)

8. Publish a web app

One line: any agent can serve a page / API / webhook and expose it on a throwaway public URL — no domain, no account. Offer to scaffold a hello-world and spin an ephemeral tunnel so they see a live link. (Depth: publish-web.)

9. Custom skills

Explain: skills are reusable playbooks at .claude/skills/<name>/SKILL.md. Ask if there's a task they do repeatedly; if so, capture it as a skill in ~/.claude/skills/ (per-agent). Built-in skills already cover the basics — list them with what's under /usr/share/nanabox/.claude/skills.

10. Wrap

Recap in ≤5 bullets what's now set up and the obvious next moves. Point at the per-feature skills for depth (manage-agents, agent-2-agent, browser, publish-web, nightly-cycle). Close: "You're set — just talk to me, or any <handle>@<box> session, to get going."

More from this repository

same repository

browser

eladb/nanabox

Drive the long-lived Chromium running on this nanabox over the Chrome DevTools Protocol. Use when the user asks to scrape a page, automate a browser flow, fill a form, log into a site, take a screenshot, or otherwise control a real browser. The user can watch and take over the same session from their own browser tab — useful for one-off logins, captchas, or MFA.

2026-06-263

publish-web

eladb/nanabox

Publish a web app (a page, an API, a webhook receiver) from this nanabox. Use when the user asks to host a page, expose an API, share a script as a URL, set up a webhook, or otherwise serve something over HTTP from this box. Your app binds a unix socket; Caddy reverse-proxies your handle's path to it, and an ephemeral cloudflared quick tunnel makes it reachable on the public internet.

2026-06-263

agent-2-agent

eladb/nanabox

List the other agents on this box and send them a message (one-to-one or broadcast). Use when the user asks "who else is running here?", "tell the other agents X", "broadcast to all agents", or when you want to coordinate work with a peer agent — e.g. handing a sub-task to a peer while you stay focused on your repo. Messages are fire-and-forget; the receiver reacts on its next turn.

2026-06-263

manage-agents

eladb/nanabox

Understand and operate the agents on this nanabox — list them, peek at a session for debugging, restart a wedged one, and push a one-off prompt in. Use when the user asks about "the agents" or a specific agent's session. Creating/removing agents and changing box config are the box operator's job, not yours.

2026-06-263

nightly-cycle

eladb/nanabox

End-of-day consolidation routine. A box-wide systemd timer fires `/nightly-cycle` overnight (around 03:00–05:00 UTC; the firing is staggered across the fleet and across agents on a box to avoid an API-rate-limit burst) for each agent on this box. The agent reviews the day's work, distills durable insights into memories and skills, writes a brief day summary at its per-handle path `~/.agents/<handle>/nightly-summary.md`, then `/clear`s the session. When fired by the timer (vs run manually), the box ALSO restarts the agent's systemd unit after the summary lands so the next session re-reads the runtime CLAUDE.md + skills — which `/clear` alone doesn't reload.

2026-06-263

name

onboard

description

onboard

You are the root/admin agent of this box (name: cat /etc/nanabox/handle). Run a hands-on onboarding that gets the user productive fast and teaches every capability.

How to run it — read this first:

Be interactive and terse. Ask ONE decision at a time (use multiple-choice / ask-user prompts), DO the thing, confirm in one line, move on.
Demonstrate, don't lecture. No walls of text. Optimize for a working box, not prose.
Let the user skip any step ("skip" / "not now"). Keep an internal checklist so you can resume if interrupted.
You have passwordless sudo (you're the box admin) — verify once with sudo -n true.

1. Orient (one line)

"I'm the admin agent for . I'll set you up and show you what this box can do — tell me to skip anything." Then go.

2. About this box → your system prompt

3. GitHub access (optional)

Ask: want this box to work with your GitHub repos?

If yes: run gh auth login and relay the device code + https://github.com/login/device URL it prints; the user authorizes in their browser. Verify with gh auth status. Agents can now clone / push / open PRs.
For tighter scope, mention they can instead pipe a fine-grained PAT or a GitHub App installation token: gh auth login --with-token.

4. Your first agent

5. Talk between agents

Demonstrate live: agents send <handle> "ping from root — reply to confirm", then agents list. Mention agents broadcast for all agents at once. (Depth: agent-2-agent.)