agentleFS
5hirish5hirish/duct/.agents/skills/new-page/SKILL.md

new-page

Skill11 starsChanged 4 days agoMITView on GitHub

Create a new audience-specific landing page following the for-*.html pattern

10 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use new-component.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • violacaseSkill

    new-component

    Scaffold a new Vue 3 component using reka-ui primitives and Tailwind CSS. Use when the user asks to create a component, UI element, widget, or any .vue file in the components/ directory.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 8 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    No testimonials, no stats band, no interactive demo.

    line 48
  2. Distinctive:

    FAQ: Write the answers in the page's voice.

    line 94

“Common” is measured: kept by at least 6 in 10 of the 8 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Creates a new for-<audience-slug>.html page for one audience. Keeps the site's design, CSS and structure; only the copy and the shots change.

/new-page for-engineering-teams "for Engineering Teams" "Stop losing signal in the noise." "Duct connects Sentry, Linear and GA4 and answers why the error rate moved."

Before writing: answer three questions

  1. What job does this audience hire Duct for? The outcome, not the feature. "Know why activation dropped before standup", not "cross-tool aggregation".
  2. What is the specific pain? Two conflicting signals in one sentence. "Search Console says impressions are up. GA4 says trials are flat."
  3. Which tools do they live in? Name them in the subheadline. It is the highest-leverage line on the page.

And the one AGENTS.md asks first: what is the hypothesis? It goes in the <!-- EXPERIMENT: --> comment after the GTM noscript.

The page, in order

Copy site/for-product-intelligence.html. Every for-* page keeps this sequence; each section earns the next.

Hero          Stop X. Start <em>Y</em>. Tools named in the sub. Download + Star on GitHub.
Shot          One README screenshot under the hero (.shot). The page's heaviest element.
Tool strip    The marquee: the tools this audience recognises.
Problem       "You're doing the right thing. But no one <em>connected it</em>." Three pain
              bullets, the disconnected-tools diagram, one "After Duct" sentence.
What it does  Three .job cards: a shot plus When / Today / With Duct.
How it works  Connect / Ask / Approve. Name the tools in step one. Keep the channel.
Who it's for  "The [role] who owns [thing] without a [resource]". Four fit cards,
              the last one "Not yet". Inline CTA: That's me — download Duct ↓
FAQ           Five or six real questions. The JSON-LD is rebuilt from the <details>.
CTA           The shared partial.

No testimonials, no stats band, no interactive demo. All three were on these pages once and all three were removed: fake quotes, invented numbers, and a mock brief that lost to a real screenshot.

Section notes

<head>. Title Duct <suffix> — <seven-word value line>, canonical https://getduct.ai/for-<slug>, og:description and twitter:description 120–140 characters and different from each other, WebPage plus FAQPage JSON-LD.

Hero. Headline ≤60 characters, <em> on the aspiration, never on the pain. No "powerful", "seamless", "all-in-one". Button copy is Download Duct ↓ at the top and the bottom, always href="/download" data-duct-download. Footnote: Free · No credit card · For [role] at [size] companies, honest on both.

Shot. The hero shot is a session asked from this audience's side. Three exist: product-session (why did activation drop), paid-session (where is the budget leaking), insights-session (why are signups down); each ends on its brief in the artifact pane, with the sidebar closed so the canvas is the picture. A new audience gets a new one: its answer in ANSWERS, its brief through brief() in the story, its stream in scripts/shots/fixtures.mjs, a scenario in scenarios.mjs, then shoot it. The job cards take focused shots, never a whole window: answer-<audience> (one question, the sources read, the answer; a new audience adds its entry to ANSWERS in the story and gets the card for free), review-card, content-plan, memory-timeline, connectors, and audit-report (the story's AUDIT_REPORT_HTML, a document like the briefs; an html scenario in scenarios.mjs needs no app to shoot). Cap .shot-frame at the image's 1x width so a 2x capture never upscales. Never draw a mock of a screen that exists.

Problem. Mirror the audience's own words. The diagram keeps the window chrome and the dim pattern; the "After Duct" line is [A] → [B] → [C] → [outcome]. One story, automatically connected.

Jobs. Three cards, three different shots. When is a moment, Today is the cost, With Duct is what it reads and what it proposes. Numbers beat adjectives.

How it works. The bodies are shared across pages on purpose; only the tools in step one change.

Who it's for. The "Not yet" card is the most trusted line on the page. Name who this is not for.

FAQ. Write the answers in the page's voice; search engines quote them. The money answer is the site's one sentence: free with your own keys, always; a paid plan later bundles the models; nothing open today moves behind it, with a link to /open-source#pricing.

Icons. Sprite only: <svg class="ic" aria-hidden="true"><use href="/assets/icons.svg#name"/></svg>. A missing name goes into scripts/build_site_icons.py.

Accent. One <style> block overriding --orange and --orange-h, nothing else. Warm for urgency (orange, amber), cool for precision (indigo, slate, teal), mid for growth (green, violet).

Done means

  • Hypothesis in the EXPERIMENT comment
  • Headline ≤60 chars, tools named in the sub, both CTAs identical
  • One shot, three job cards, no mock of a screen that exists
  • "Not yet" card names who it is not for
  • No emoji, no stats band, no testimonials, no demo
  • Added to site/sitemap.xml; make i18n run so site/<lang>/ has the page in every language (the generator adds those sitemap entries); make check-site green
  • Reviewed at 1440 and 390 px: two-line hero, nothing horizontal

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.