Shell Tool · engineering · Sentinel

Clone a Website

Fire ultron-capture: spider the site, deploy a branded clone, surface the URL when ready.

Updated today

View as MarkdownsentinelsonnetstandardBackgroundMax 2 turns

Overview

Long-running async clone via the ultron-capture pipeline. Call the start_site_capture chat tool with the URL. It returns IMMEDIATELY with a job_id — the user sees the job in the Background Jobs panel. Simple sites finish in ~30s, full SPAs (Apple-scale) take up to 40 min. When done, bg_trigger fires and you'll resume the conversation with the branded deploy URL (<slug>.clones.51ultron.com) and the slug. From there: if the user wants to edit, shell_exec wget the deployed site into /work/<slug>/ and use the edit-html-clone skill for surgical edits, then ultron-deploy to ship the modified version. END THE TURN immediately after calling start_site_capture — do NOT poll, do NOT loop, do NOT use the in-container ultron-clone (that's the broken v1 that misses SPA content).

When to use this

user wants to clone an existing website
user says 'rebrand <url>' / 'make a copy of stripe.com' / 'mirror linear.app'
user wants to swap copy on a competitor's landing and ship a variant
user wants the site as a starting point for edits + deploy to Pages

When NOT to use this

user wants to scaffold a NEW web app from scratch → use spin-up-web-app
user wants to scrape data only (no rebuild/deploy) → use scrape_website
user wants only a screenshot of a page → run ultron-screenshot via shell_exec

Example prompts

clone https://linear.app and rebrand it for our launch

make a copy of stripe.com landing page

mirror flora.ai so I can edit the copy

rebrand vercel.com with our colors and ship to Pages

Inputs and output

Inputs

Field	Description
`url`	the URL to capture
`branch`	optional kebab-case branch name for the Pages deploy

Output

job_id (visible in Background Jobs panel). Branded URL surfaces via bg_trigger when capture completes (30s–40min depending on site size).

Runtime profile

What the engine commits when this skill runs.

Property	Value	Meaning
Model tier	`sonnet`	The balanced default model class. Trades quality against cost for the vast majority of skill runs.
Cost class	`standard`	The balanced default model. Right for most skills.
Turn budget	`2`	Hard cap on tool-calling iterations before the engine forces a final answer.
Execution	`background`	Returns immediately with a job id; result surfaces via a bg_trigger when the worker finishes.

Under the hood

Tools the engine exposes to this skill and integrations it needs.

Resource	Kind
`start_site_capture`	tool

Tags: engineering, clone, deploy, async, background

Execution template

Shell tools dispatch via a command template. The model substitutes user inputs into the placeholders before firing the underlying tool.

start_site_capturebash

1call start_site_capture({{url}}, {{branch}}) and END THE TURN. bg_trigger will resume when the capture is done.

Related skills

Skill	What it does
Spin up a Web App	Scaffold a fresh web app in /work, start the dev server, surface the live preview URL.
Build an Agent	Spin up an autonomous agent on the agent-runtime satellite — try templates first, custom build only if no fit.
Deploy to Cloudflare	Ship the current /work to Pages, Workers, or Containers — then verify the live URL.
Push to GitHub	Initialize /work as a git repo, create the GitHub repo, push, return the URL.
Swap Page Sections	Reorder <section> blocks in a minified HTML file with byte-level safety.
Refactor Across Files	Multi-file string operations with quote and escape safety — uses bun -e, not sed.

Invoking this from an agent

Three paths reach this skill. From the chat UI, a user can type the persona slash command followed by a natural request and the discovery step resolves to this skill automatically. From the MCP server, fetch the skill detail with get_skill({id: "clone-website"}) and then invoke it through the agent runtime once the authenticated tier ships. From your own code, hit /docs/skills/clone-website/llm.txt for the token-efficient markdown body and feed it to your model directly.

Note

Every skill page has a canonical permalink and a markdown alternate that LLM crawlers consume via Accept: text/markdown. The full machine-readable catalog lives at /.well-known/agent-skills/index.json.