Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Animated Marketing Website from a Business Brief (Astro)

    1

    Turn a business brief into one paste-ready prompt that builds a tested, design-graded Astro marketing site with animation components, unattended.

    $19

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Animated Marketing Website from a Business Brief (Astro)

    Animated Marketing Website from a Business Brief (Astro)

    Example session with this skill installed

    Use the tier2-website-pipeline skill. Business: a florist in a residential district of a European capital. Google Maps rating 4.8 with 66 reviews, 8 years open, no website, Instagram and Facebook only. Services: bouquets, event arrangements, weekly subscriptions. Phone +40 7XX XXX XXX, WhatsApp orders. Language: Romanian. Goal: a 4-page pitch demo (home, gallery, events, contact) I can show the owner next week.

    • Read your context and instructions
    • Compiled the animated marketing website

    Summary

    Excerpt from a real run; business details redacted.

    Tier line: TIER 2 — polished/marketing. Local florist, mobile Maps/Instagram traffic, needs premium feel without 3D.

    Direction: Editorial Organic. Palette #FAF6EF paper / #2A2E26 ink / #3E5934 moss (dominant) / #C4622D terracotta (spark only) / #EDE4D3 sand. Fraunces display + Karla body via @fontsource.

    The skill then emitted one 280-line prompt. Pasted into Claude Code, the unattended run produced:

    pipeline/spec.md            faithful expansion, no scope additions
    pipeline/ownership.md       7 workers, exclusive file lists
    pipeline/art-direction.md   photo manifest, one grade recipe, per-page archetype map
    pipeline/build-notes.md     ASSUMPTIONS-MADE: 4 · component swaps: 0
    pipeline/test-report.md     6 checkers, evidence pasted
    pipeline/design-review/     24 screenshots (4 pages x 3 widths x 2 motion modes)
    pipeline/audit-log.md       7 checks PASS, 1 FLAG resolved
    pipeline/review-verdict.md  APPROVED
    

    Visual-craft table from test-report.md

    PageGradeNote
    indexPASSAsymmetric hero, moss-dominant, terracotta only as spark
    galeriePASSGenuine asymmetric masonry, duotone tiles avoid beige-on-beige
    evenimentePASSEditorial zigzag, one decisive image per service
    contactSLOP → PASSPhone pill collided with the social row at 1440 and 375; routed back to its owner, re-graded with overlap = 0

    Reviewer verdict excerpt

    4 pages build and render, 4 routes, sitemap emitted, 0 original images leaked into dist. 8 React Bits components installed for real (gsap, motion, canvas sources present), every usage mounted client:visible. @theme carries all five hexes verbatim. Fonts self-hosted, no Google CDN. WhatsApp link decoded to the exact template text with diacritics intact. MATCHES THE PLAN — YES, two minor non-blocking accuracy defects listed in the punch-list.

    Next steps the run reported

    • Replace the 6 // FAKE-REPLACE testimonials with real quotes
    • Confirm 3 [CONFIRM PRICE] markers with the owner
    • Swap placeholder photos
    • Deploy was out of scope by design: one static-host command afterwards
    Visual-craft table from test-report.md

    Connects securely to your tools. The creator never sees your data.

    About this skill

    What you get. A built static marketing site and a pipeline/ folder holding the spec, a test report, a PASS or SLOP design grade for every page, and a reviewer verdict. In the demo run: 4 pages, 24 screenshots, verdict APPROVED.

    How it runs. You give it a business (Maps listing, Facebook page, phone number). It researches the business, picks one of 11 named design directions (hex and OKLCH palette, self-hosted font pair, motion vocabulary), and writes one long prompt. Paste it into Claude Code and a four-agent pipeline runs: planner, coder (5 to 20 parallel workers with a file-ownership map), tester (six checkers, screenshots at 375, 768 and 1440), reviewer. Max two rejection loops, then it halts with a failure report. Typical run: 25 to 35 minutes. Anything unverified ships as a // FAKE-REPLACE or [CONFIRM ...] marker. Example use: a pitch demo for a local business.

    Stack. Astro static output, Tailwind v4, React islands only for registry animation components (React Bits first), @fontsource, astro:assets.

    What's in the package.

    • SKILL.md (347 lines)
    • reference/ (7 files, ~1,600 lines)
    • agents/: 4 agent definitions plus install notes
    • README with an example brief and the output file list

    Honest limits. No deployment, no backend, no CMS. Real prices, reviews and photos are owner-supplied and ship as marked placeholders. Requires the free, public impeccable skill; the pipeline halts without it. Runs on Claude Code; other SKILL.md agents can read the protocol but the four bundled agents are Claude Code agent files.

    Requires impeccable craft skill. No backend/CMS/payments. Unattended runs require specific agent folder setup. High token usage per build run.

    How to use

    1. 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. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 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 23 days ago

    Requires

    shadcn
    Node.js
    react-doctor
    astro
    wrangler

    Permissions

    Browser
    Network Access
    Read Files
    Write Files
    Listed23 days ago
    Updated17 days ago

    What's inside

    Frequently Asked Questions

    This skill eliminates 'AI slop' by automating deep business research and generating a multi-agent orchestration prompt that enforces high-end design tokens and specific React registry components.

    The skill is optimized for Claude Code sessions, as it generates a complex, multi-role prompt (Planner, Coder, Tester, Reviewer) designed for autonomous terminal-based builds.

    You will receive the 'tier2 website pipeline' skill logic that performs the research and prompt engineering; however, you will need a separate 'impeccable' design-craft skill for the visual enforcement layer.

    No, this pipeline focuses on high-fidelity frontend delivery using Astro 5 and Tailwind v4; it does not currently support backend logic like databases, payment gateways, or CMS integrations.

    Yes, the pipeline includes a decorative JS budget enforcement of 60KB and logic for visual regression and accessibility checks to ensure the site is production-ready.

    New in Frontend & Web Apps