- Home
- Skills
- Frontend & Web Apps
- Landing Page Designer: Your AI Builds Pro Pages
More screenshots
Works with the AI Agents you already use
Landing Page Designer: Your AI Builds Pro Pages
Turn a one-line brief into a responsive landing page with a real design system, then check it in light and dark mode at three widths until nothing is broken.
$19
About this skill
Your model can design a landing page that looks finished. Then the OS switches to dark mode and the nav button disappears (1.01:1), or every brand-teal button fails contrast (3.74:1). You cannot see that in a screenshot.
Measured on 2 plain-model pages: render_check found 35 and 36 failing text runs. One fix loop later (6 and 4 CSS edits, same design): 0 FAIL in light and dark at 375, 768 and 1280 px.
Your model designs, the studio checks, the model fixes. Three local tools:
- render_check.py loads any page in a real browser at three widths, light and dark. It measures sideways scroll and names the element that breaks out. It checks every text run against what is really painted behind it, after opacity, filters,
::before/::aftertext and placeholders. It also checks tap targets (WCAG 2.5.8), headings, alt text and labels, and names the selector and colours to fix. Without Playwright it does a static read and says what it skipped. - brief_to_tokens.py turns one brand colour into a palette, type scale and 8-pt spacing for light and dark, with all 22 text pairs at AA. Mid-tone brands keep white button text on a step darker, not muddy dark text.
- scaffold.py (optional) gives a fast clean baseline from tested sections. The hero visual is your content: a list of real rows, metrics, a chart that can go down, code, or your own HTML. Copy lint flags hype words, vague buttons, placeholders and unproven proof.
Tested on public pages (commits pinned in the docs):
| Page | Result | Worst contrast |
|---|---|---|
| startbootstrap-landing-page | 3 FAIL | 15 runs, 4.27:1 |
| startbootstrap-agency | 3 FAIL | 27 runs, 1.55:1 (white on yellow) |
| tailwindtoolbox Landing-Page | 2 FAIL | 6 runs, 3.53:1 |
Every flagged ratio was recomputed independently and matched.
Honest notes: both plain pages were well designed. The studio does not design for you; it makes your model's design ship-quality. n = 2 briefs. A browser check takes about 5 to 40 s, mostly browser start-up.
What it does not do: draw illustrations, or test clicks, menus, hover or form states. It cannot measure text over images or gradients: it counts those and asks for a check by eye. It cannot tell an invented claim from a real one, so the instructions tell the agent to use only facts from your brief.
Works in Claude Code, Cursor, Codex CLI, Copilot and any SKILL.md agent.
Private by design: standard-library Python on your machine, no telemetry. During checks the network is blocked unless you pass --allow-network.
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 any payment.
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 27 days ago
Requires