- Home
- Skills
- Mobile Development
- Every Screen — App & Site Designer That Works on Phone and Desktop, Proven
More screenshots
Works with the AI Agents you already use
Every Screen — App & Site Designer That Works on Phone and Desktop, Proven
Your agent plans the user's task first, then designs and builds it for phone and desktop, remembers your design system, and proves it works with real browser tests.
$9.99
Every Screen — App & Site Designer That Works on Phone and Desktop, Proven
Example session with this skill installed
Build an orders admin for our shop staff. They need to find problem orders fast and fix failed payments, on desktop and on their phones.
- Read your context and instructions
- Compiled the every screen
Mode: standard. Writes .everyscreen/brief.json and task contracts for find-order and resolve-failed-order (states: loading, load-error, view-only, failed, retrying, retry-failed, paid, cancelled). Retrieves 8 matching rules (e.g. orders.attention-queue, orders.failure-reason, records.list-detail-route) plus always-on accessibility rules. Builds a 'Needs attention' view, plain-language failure reasons, an idempotent retry, a cancel dialog that names the refund/restock consequence, and three layouts chosen by content fit (list < 960px, sidebar + table, split pane >= 1360px). Ends with verification.json: 18 checks pass, screen-reader pass marked not-run with the reason.
Connects securely to your tools. The creator never sees your data.
About this skill
Most AI-made interfaces look fine in one screenshot and fall apart on a phone, in an error state, or the next time you ask for a change. Every Screen gives your agent a method instead of a mood board.
- Task first. Before any styling, the agent writes a task contract: who acts, on what, toward which outcome, through which states (loading, empty, error, no permission, plus real domain states like "payment failed").
- Adapts the interaction, not just the grid. Tables, filters, list/detail, navigation and primary actions change in documented ways between phone and desktop, and the same task is checked at both sizes. Breakpoints come from content fit, not device names.
- Remembers your design. Tokens (light + dark), decisions and screens live in
.everyscreen/in your repo, so later sessions continue the same design. Change a token andimpactlists every screen to re-check. - Shows its evidence. Every check is recorded as pass, fail, blocked or not-run with what was actually observed. No "looks great!" without proof.
- 45 original guidance records across web, iOS, Android, desktop apps and five product types (SaaS/admin, analytics, storefront, merchant ops, marketing/content). Each says whether it is a WCAG requirement, platform guidance or a heuristic, and how to verify it.
- Zero-dependency CLI (Python standard library): validate briefs, retrieve the right rules, export tokens to CSS/TypeScript, check contrast for every token pair, and find change impact.
- Three modes, so a label fix stays quick and a new product gets a real discovery pass.
Built and tested on an original shop + merchant admin (React + TypeScript): 72 browser checks pass from 320 to 1920 px in light and dark themes, including keyboard paths, automated accessibility scans and layout-fit checks.
Support levels, stated honestly: desktop and mobile web with React + TypeScript are implemented and tested. Other web stacks get design guidance. Native iOS, Android and desktop apps get specification-level handoffs, not verified native code.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Mobile Development.
Native iOS/Android/desktop output is specification-only (no verified native code). Only React + TypeScript web is implemented-tested; Vue, Svelte and Angular get guidance only. Automated accessibility checks cover part of WCAG; screen-reader testing is not performed by the skill. Comparative benchmarks against other skills have not been run, so no superiority claims are made.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 1 day ago
Nothing to install. It runs in your AI agent.
- Optional
Python 3.11+
Browser toolorPlaywright
Python 3.11+ (standard library only) powers the optional CLI for rule matching, token export, contrast and impact checks. A browser tool enables screenshot and interaction checks. Without them the method still works and unrun checks are reported as not-run.
Permissions
File Scopes
The CLI makes no network calls. Export never overwrites a different file without --force and refuses paths outside the export folder.