AI Summary Request

Overview

This skill creates a footer component with AI platform icons that allow visitors to request AI summaries of a website. The workflow begins by detecting and confirming the website domain, then asks whether to optionally generate an llms.txt file for enhanced AI discoverability.

When to Use This Skill

Use this skill when:

• User wants to add AI platform integration buttons to their site
• User wants visitors to easily get AI summaries of their business
• User mentions "AI summary" or AI platform buttons
• User wants to create an llms.txt file for their website (optional)
• User asks about making their website AI-discoverable

Workflow

PREREQUISITE: Domain Detection and Confirmation

Before proceeding with any steps, you MUST determine the website domain:

Step 1: Attempt automatic detection

Search for domain information in this priority order:

1. Environment files (.env, .env.local, .env.production) - look for SITE_URL, BASE_URL, NEXT_PUBLIC_URL, NUXT_PUBLIC_SITE_URL, VITE_SITE_URL, or similar
2. Deployment configs (vercel.json, netlify.toml, firebase.json, fly.toml) - check for domain/alias settings
3. CNAME files in public/ or root directory
4. Configuration files (package.json homepage field, next.config.js, nuxt.config.js, vite.config.js)
5. HTML files - meta tags or <base> tags in index files
6. README or documentation mentioning the live site URL

Step 2: Confirm or request domain

If a domain is detected, ask the user to confirm or provide an alternative:

"I detected that this site appears to use the domain [DETECTED_DOMAIN]. Is this correct? If not, please provide the correct domain."

If no domain can be detected, ask the user directly:

"I couldn't automatically detect the domain for this website. What domain should be used for the AI summary links? (e.g., https://example.com)"

Step 3: Normalize and store the domain

• If the user provides a domain without a protocol, prepend https://
• Remove any trailing slashes
• Store the normalized domain as ${WEBSITE_URL} for use in subsequent steps

Example normalizations:

• example.com → https://example.com
• https://example.com/ → https://example.com
• http://example.com → keep as-is (user may have specific requirements)

---

PREREQUISITE: llms.txt Decision

After confirming the domain, ask the user:

"Would you like me to also generate an llms.txt file for this website? This file helps AI models better understand your site's content and structure, which can improve the quality of AI-generated summaries.

• Yes: I'll create a curated llms.txt file and the footer component will reference it
• No: I'll create just the footer component, and the AI prompt will direct users to explore the website directly"

Store the user's choice as ${INCLUDE_LLMS_TXT} (true/false).

---

STEP 1: Generate llms.txt File (OPTIONAL - Only if user chose "Yes")

Skip this step entirely if the user chose not to generate an llms.txt file.

Generate a high-quality llms.txt file for a given website that helps language models quickly understand and reference the site's most valuable content.

Analyze the Website's Content Hierarchy

• Inspect any available XML or HTML sitemaps (product, page, post) and explore navigation menus to identify major product categories, services, key informational pages, and blog or news posts
• Focus on top-level categories and representative subcategories or key products, rather than listing every single URL
• Identify important support or policy pages (e.g., About, Contact, Terms, Privacy, Shipping & Returns)

Select and Curate Content

• For each major category, select a primary landing page and a few representative subcategories or high-value product pages
• Include links to authoritative guides, how-to articles, calculators, training programs or other resources
• Choose a handful of blog posts or case studies that showcase different topics (tutorials, industry insights, success stories). Avoid listing dozens of posts
• Exclude low-value pages such as checkout flows, login pages, or marketing fluff

Draft Clear Descriptions

• Use concise, factual descriptions (10-15 words) that explain what each page covers
• Avoid hype or redundant repetition of the page title
• When summarizing subcategories, group them in parentheses for brevity

Build the llms.txt in Markdown

• Start with an H1 title containing the website or company name and a short descriptor
• Add a blockquote (prefaced by >) that succinctly describes what the business does and who it serves
• Optionally include a sentence explaining that the list is curated for AI consumption
• Organize content with H2 headings such as:

  • Product categories (or Services if it's a service business)

  • Representative pages or ## Popular products

  • Guides & resources

  • Articles & case stu

...