Setup Links

The "setup-links" skill constructs or updates a product's "/links" page, adhering to a predefined RDF standard for link-in-bio pages. It ensures the page is static, noindexed, includes privacy-preserving click tracking, and integrates necessary legal footers and internationalization.

What it does

Creates a dedicated "/links" page, which serves as a product's link-in-bio page for social profiles.

This page follows a strict organization-wide standard, ensuring it is static, rendered without site chrome, and configured with "noindex,follow" directives.

It implements PII-free click tracking for user interactions, includes logic for sponsored links, and generates i18n keys for all locales.

The skill integrates necessary imprint and privacy policy links into the page footer.

When it fits

  • When a new product requires a standard-compliant link-in-bio page for its social media presence.

  • When an existing link-in-bio page needs migration to meet the organization's RDF standard, including specific button configurations and tracking.

  • When ensuring all product link-in-bio pages are consistent in appearance, functionality, and tracking across the organization.

Setup

  1. Copy the "setup-links" definition file to ".claude/skills/setup-links/setup-links.md" in the repository.

  2. Invoke the skill by typing "/setup-links" in Claude Code. The agent will then guide the setup process.

The content

Setup Link-in-Bio Page

Build this product's /links page in the one form every RDF web product uses: the page the social profiles link to.

Contract reference: load .claude/ref/growth/link-in-bio.md first. It is the single source of truth — this skill only executes it. Reference implementation: padel-league.richi.solutions.

Context probes:

Monorepo web app: ![ -f apps/web/package.json ] && echo "apps/web" || echo "none (single app)" Framework: !for d in apps/web .; do [ -f "$d/package.json" ] || continue; grep -q '"next"[[:space:]]*:' "$d/package.json" && echo "next ($d)" && break; grep -q '"vite"[[:space:]]*:' "$d/package.json" && echo "vite ($d)" && break; done App Router dir: !ls -d apps/web/src/app apps/web/app src/app app 2>/dev/null | head -1 || echo "none" Existing link page: !find app src/app apps/web/app apps/web/src/app src/pages -path '*links*' -name 'page.tsx' 2>/dev/null | head -5; grep -rln '"/links"' src 2>/dev/null | head -5 Locale segment: !ls -d app/\[locale\] src/app/\[locale\] apps/web/app/\[locale\] 2>/dev/null || echo "single-locale" Locales: !ls src/i18n/locales apps/web/src/i18n/locales 2>/dev/null | tr '\n' ' ' Analytics: !grep -o '"@vercel/analytics"' package.json apps/web/package.json 2>/dev/null | head -1 || echo "no @vercel/analytics" Funnel registry: !ls src/domain/funnels.ts src/lib/funnel.ts 2>/dev/null || echo "none" Templates: !ls .claude/skills/setup-links/templates/next 2>/dev/null && echo "present" || echo "MISSING — sync .claude first"


Step 0: Eligibility

Stop and report if this repo has no public web surface (mobile-only .app. repos, backend/automation repos, behind-auth-only tools). Work on a feature branch in a worktree (/new-worktree feat/links-page). In a monorepo every path resolves under apps/web/.

Step 1: Inventory

Read and note before writing:

  1. An existing link page (probe above) or third-party link page (Linktree etc. — ask the user for the current bio URL if unknown). Record its buttons and order; a migration keeps them unless the user says otherwise.
  2. The repo's helpers for: server-side translation, page metadata, locale routing (localizedPath/asLocale), the analytics facade, the funnel registry (if any).
  3. Imprint and privacy page paths (legal-pages.md); stop and report if either is missing — the page must link both.
  4. The brand logo (a ≤ 96 px WebP; create one from the existing logo with sharp if absent) and the brand colour tokens (Tailwind theme).

Step 2: Decide the Buttons

Propose to the user, in one message: 3–6 buttons (id, label, destination), which one is highlighted and why, and the order — by business value (contract §4). Wait for confirmation before writing. Mark affiliate destinations as sponsored.

Step 3: Copy the Templates (Next)

Copy every file from .claude/skills/setup-links/templates/next/ to the same relative path (drop .tmpl; app/ → the App Router dir; without a [locale] segment use app/(bio)/links/page.tsx). Then adapt:

TemplateMap to
@/i18n/server getServerTthe repo's server translation helper
@/lib/seo/metadata buildPageMetadatathe repo's metadata builder (canonical + hreflang); else a plain Metadata object with alternates.canonical
@/lib/seo/locale-routesthe repo's locale routing; single-locale: drop params/locale, use plain paths
@vercel/analytics trackthe repo's analytics facade
@/lib/funnel funnelStepkeep if the repo has a funnel registry; otherwise remove funnelStep, BioPageView and the funnel assertions in the test
__PRODUCT_NAME__the product name
bg-primary / bg-accent …the repo's brand tokens

Fill ITEMS, HIGHLIGHT, SPONSORED and BioTarget from Step 2, and replace the TODO(setup-links) comment with the reasoning for the order. Do not change the mechanics (static, noindex/follow, structure order, click event shape, sponsored rel). A deviation needs a documented reason in the file.

Vite exception repos: build the same structure as a route rendered outside the app layout, with <meta name="robots" content="noindex,follow">; reuse BioLink.tsx without "use client".

Step 4: Wire It

  1. Funnel registry (if present): add the links funnel — step view, one outcome per button id (contract §5).
  2. i18n: add the links.* keys (contract §4) to every locale file; run the repo's key-parity test.
  3. Routes: /links stays out of PUBLIC_ROUTES and out of EXCLUDED_PREFIXES. If the repo keeps a list of page paths for locale routing (e.g. content routes), add /links there.
  4. Middleware / chrome: confirm the route group renders no site header, footer, nav, chat widget or banner other than the consent banner.
  5. Remove the legacy: delete an old link page and redirect its URL to /links in next.config redirects().

Step 5: Verify

  1. npm run lint, npm run typecheck, npm run test, npm run build — green; the build lists /links as static (○ or ●).
  2. npm run start, open /links (and one other locale) at 390 px width: structure order, one highlight, every button reaches its destination, imprint/privacy present, <meta name="robots" content="noindex, follow"> in the HTML, no site chrome.
  3. Walk the contract checklist (§8) and report each item.

Final Step: Commit and Hand Over

Commit with a body, e.g.:

feat: add the link-in-bio page at /links

Static /links page for the social profiles, outside the site chrome and
noindexed: <n> buttons led by <highlight>, clicks reported as
bio_link_click and in the links funnel without PII. Follows the
org-wide link page standard (ref/growth/link-in-bio.md).

Report to the user: the buttons in order, and that after deploy the bio of every social profile must point to https://<product-domain>/links.

Served straight from the framework this site runs on. This page always shows the current version.

The same tools, a niche of its own.

The same tools run PadelLeague, MovieMind and Hookr. How Richi uses them to build products for a niche is described under partnerships.

View partnerships