Aeo Llmstxt

AEO llms.txt — LLM-Readable Site Manifest Generator

Description

The llms.txt sub-skill generates an llms.txt file for your website following the emerging standard for communicating site structure, content hierarchy, and preferred citation format to AI engines. Think of llms.txt as robots.txt for the AI age — while robots.txt tells crawlers what they can access, llms.txt tells AI engines what they should know about your site and how they should cite it.

The llms.txt standard is still emerging but gaining adoption. Early adopters gain a structural advantage as AI engines increasingly look for machine-readable site metadata beyond traditional sitemaps and schema markup.

Command

/aeo llmstxt  [--output ]

The llms.txt Standard

Purpose

llms.txt provides AI engines with a concise, structured overview of a website that includes:

1. Site identity — What the organization is, what it does, who it serves
2. Content map — Key pages organized by category with brief descriptions
3. Expertise signals — Topics the site is authoritative on
4. Citation preferences — How the site prefers to be cited and attributed
5. Contact and verification — How AI engines can verify information or request corrections

File Location

The file should be placed at the root of the domain:

https://example.com/llms.txt

Format Specification

The llms.txt file uses a simple, parseable plain-text format with section headers and structured entries:

# Site Name
> Brief site description (1-2 sentences)

## About
- Organization type: [Company/Nonprofit/Publication/Individual]
- Industry: [Primary industry]
- Founded: [Year]
- Location: [City, State/Country]
- Expertise: [Comma-separated list of topics]

## Key Pages
- [Page Title](https://example.com/page): Brief description of page content
- [Another Page](https://example.com/another): Brief description

## Content Categories
### [Category Name]
- [Page Title](URL): Description
- [Page Title](URL): Description

### [Another Category]
- [Page Title](URL): Description

## Citation Preferences
- Preferred name: [How you want to be referred to]
- Preferred URL: [Canonical domain]
- Attribution format: [How to attribute]

## Contact
- Corrections: [email for factual corrections]
- Press: [press contact]
- Website: [main URL]

## Verification
- Official domain: [domain]
- Social profiles: [list of verified social URLs]

Generation Process

Step 1: Site Discovery

The generator crawls the target site to identify:

• Homepage title and meta description
• Navigation structure (main menu items)
• Key landing pages (linked from homepage)
• Blog/resource section structure
• About/team pages
• Product/service pages
• Contact information
• Schema.org Organization markup
• Existing sitemap.xml content

Step 2: Content Classification

Each discovered page is classified into categories:

• Core Pages — Homepage, About, Contact, Pricing
• Products/Services — What the organization offers
• Resources — Blog posts, guides, whitepapers, tools
• Authority Content — Case studies, research, data, expert content
• Support — FAQ, documentation, help center

Step 3: Expertise Extraction

The generator identifies topic expertise from:

• Blog post topics and categories
• Service/product descriptions
• About page claims and credentials
• Schema.org markup (knowsAbout, expertise)
• Title tags and meta descriptions across pages

Step 4: Citation Preference Generation

Based on site analysis, the generator recommends:

• Preferred name — Legal name vs. brand name vs. abbreviation
• Preferred URL — www vs non-www, https canonical
• Attribution format — "According to [Brand]" vs. "Source: [Brand] ([URL])"
• Author attribution — Individual authors vs. organization

Step 5: File Generation

The final llms.txt file is generated, validated, and output.

Output Format

The skill produces two outputs:

1. The llms.txt File

# Example Corp
> Example Corp is a B2B SaaS platform helping mid-market companies automate their supply chain operations. Founded in 2019, serving 500+ customers across manufacturing and logistics.

## About
- Organization type: Company
- Industry: Supply Chain Technology
- Founded: 2019
- Location: Austin, TX
- Expertise: supply chain automation, inventory management, demand forecasting, logistics optimization

## Key Pages
- [Supply Chain Automation Platform](https://example.com/platform): Core product overview with features and capabilities
- [Demand Forecasting](https://example.com/features/forecasting): AI-powered demand forecasting reducing stockouts by 40%
...

2. Implementation Report

## llms.txt Generation Report

### File Statistics
- Total pages indexed: [N]
- Pages included in llms.txt: [N]
- Content categories: [N]
- Expertise topics: [N]

### Implementation Steps
1. Upload `llms.txt` to your site root (https://yourdomain.com/llms.txt)
2. Add reference in robots.txt: `# LLMs: /llms.txt`
3. Add `` to your HTML head

### Validation
- File format: [Valid/Issues]
- URL accessibility: [All accessible/N broken]
- Content coverage: [% of site covered]

Best Practices

Content Selection

• Include pages with unique, authoritative content — not thin landing pages
• Prioritize pages with original data, research, or expert insights
• Include cornerstone content that defines your expertise
• Exclude duplicate content, paginated archives, and tag pages
• Limit to 50-100 most important pages for clarity

Description Writing

• Keep page descriptions to 1 sentence, under 150 characters
• Focus on what makes the page uniquely valuable (not generic descriptions)
• Include specific numbers or claims when possible
• Use active voice and clear language

Maintenance

• Update monthly or when major content changes occur
• Add new cornerstone content as it is published
• Remove outdated pages that are no longer maintained
• Review citation preferences annually

Integration

This sub-skill is part of the AEO toolkit but does not directly contribute to the AEO Visibility Score (it is an optimization action, not a measurement). It is typically run after an initial /aeo audit identifies opportunities for improved AI engine communication.

The Python implementation is available at scripts/llmstxt_generator.py.

A template is available at templates/llmstxt-template.txt.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: 0-shiv
• Source: 0-shiv/secondstep-claude-skills
• License: MIT

Install and usage instructions live in the source repository linked above.