llms-txt-manager

name	llms-txt-manager
description	Comprehensive system for managing llms.txt and my-blog.json files across the project. Handles creation, updating, and synchronization of AI-readable content catalogs. Automatically detects when working with llms.txt files, my-blog.json files, or updating blog content. Includes utilities for fetching latest content from query-index.json and organizing by last modified dates. Keywords include llms.txt, my-blog.json, blog catalog, content index, query-index.json, AI content discovery, update llms, update blog, content synchronization. Use when this capability is needed.
metadata	{"author":"ddttom"}

llms.txt Manager Skill

Purpose

Manage llms.txt and my-blog.json files across the project, ensuring AI systems have comprehensive, up-to-date content catalogs. Automates synchronization between blog content (query-index.json) and llms.txt files.

When to Use

This skill activates when:

• Creating or editing llms.txt files
• Creating or editing my-blog.json files
• User mentions updating blog catalogs
• User wants to synchronize content indexes
• Working with query-index.json data
• Creating folder-specific content indexes

Core Concepts

File Structure

llms.txt - AI-readable content catalog

• Location: Root or folder-specific (e.g., blogs/ddt/integrations/llms.txt)
• Purpose: Comprehensive list of content for AI systems
• Format: Markdown with metadata and categorized links
• Source: URLs from my-blog.json, text/descriptions from existing llms.txt

my-blog.json - Structured blog data

• Location: Same directory as llms.txt
• Purpose: Machine-readable catalog of blog posts
• Format: JSON with categories, posts, metadata
• Source: query-index.json filtered by folder context

query-index.json - Master content index

• Location: https://allabout.network/query-index.json
• Purpose: Single source of truth for all site content
• Contains: All pages with URLs, titles, descriptions, lastModified dates
• Updated: Automatically by the site

Folder Context

When llms.txt is in a subfolder (e.g., blogs/ddt/integrations/):

• The final folder name indicates intent/scope
• my-blog.json should exist in same folder
• Content should be filtered to relevant posts only
• Example: /integrations/ folder focuses on EDS integration tutorials

Update Logic

lastModified Date Comparison:

1. Check lastModified date in llms.txt metadata
2. Fetch query-index.json
3. Filter entries with lastModified > llms.txt date
4. Add only new/updated content

File Templates

llms.txt Template

# [Project/Section Title]

[Brief description of what this content covers]

**Last updated:** [Month Year]
**Authors:** [Author Names]

## [Category Name]

### [Subcategory Name]

- [Title](URL): Description
- [Title](URL): Description

## Version Information

**Version:** [version] (Updated: [date])
**Comprehensive catalog:** [count] posts across [n] categories
**Categories:** [List with counts]

my-blog.json Template

{
  "metadata": {
    "last-updated": "YYYY-MM-DD",
    "scope": "folder-name-or-site-wide"
  },
  "mostVisited": [
    {
      "title": "Popular Post Title",
      "url": "/path/to/post",
      "description": "Post description"
    }
  ],
  "categoryMap": [
    {
      "id": "category-id",
      "name": "Display Name",
      "count": 0,
      "description": "Category description"
    }
  ],
  "categories": [
    {
       
       
       
        
           
           
           
           
        
      
    
  

IMPORTANT: Do NOT include a latestPosts array in my-blog.json files. The view-myblog block automatically generates it from the 3 most recent posts. Including it manually requires maintenance and defeats the purpose of auto-generation.

Key Concepts

Workflows

Creating New llms.txt

Automatic Creation: The sync script now automatically creates llms.txt files if they don't exist:

1. Detects Missing Files

   • Finds all my-blog.json files
   • Checks for paired llms.txt in same directory
   • Creates missing files automatically

2. Builds Complete Structure

   • Creates header with title, description, authors
   • Builds organized sections from my-blog.json categories
   • Uses folder name for context (e.g., "ai" → "Ai Resources")

3. Populates Content

   • Fetches from query-index.json
   • Filters by folder context if applicable
   • Formats posts with titles, URLs, descriptions
   • Updates metadata to current date

Manual Creation: If you need to manually create one:

1. Determine Scope

   • Root llms.txt: Site-wide catalog
   • Folder llms.txt: Section-specific (use folder name for context)

2. Check for my-blog.json

   • Look in same directory as llms.txt
   • If missing, create one first
   • Filter query-index.json by folder path if folder-specific

3. Structure Content

   • Use folder name to determine focus (e.g., /integrations/ = EDS integrations)
   • Organize with subsections for better scannability
   • Include all relevant posts from my-blog.json
   • Add descriptive text for context

4. Add Metadata

   • Version number
   • Last updated date
   • Category counts
   • Access guidelines (optional)

Updating Existing llms.txt

1. Read Current File

   • Extract last-updated date from metadata
   • Note current structure and organization

2. Fetch New Content

   • Load query-index.json
   • Filter by lastModified > current date
   • Filter by folder context if applicable

3. Merge New Posts

   • Add new posts to organized sections
   • Update version number and date
   • Maintain text patterns and formatting

4. Update my-blog.json

   • Add new posts to appropriate categories
   • Update metadata.last-updated
   • Increment category counts

Creating my-blog.json from query-index.json

Automatic Creation: The sync script now automatically creates my-blog.json files if they don't exist:

1. Detects Missing Files

   • Finds all llms.txt files
   • Checks for paired my-blog.json in same directory
   • Creates missing files automatically

2. Initializes Structure

   • Creates empty categoryMap with all 6 standard categories
   • Sets metadata with scope (folder name or 'site-wide')
   • Sets last-updated to 2020-01-01 (extreme past, will populate on next update)

3. Populates Content

   • Fetches from query-index.json
   • Filters by folder context if applicable
   • Categorizes posts automatically
   • Updates metadata to current date

Manual Creation: If you need to manually create one:

1. Fetch Data

   curl https://allabout.network/query-index.json
   

2. Filter by Context

   • Root my-blog.json: All content
   • Folder my-blog.json: Filter by path prefix

3. Categorize Posts

   • Group by topic/category
   • Count posts per category
   • Create categoryMap

4. Structure Output

   • metadata section with date and scope
   • categoryMap with counts
   • categories with full post arrays

Command Usage

/update-llms

Finds all llms.txt files in the project and updates them with new content:

/update-llms

Or use the script directly:

node scripts/sync-blog-content.js --target=llms

Process:

1. Find all existing llms.txt files: **/*llms.txt
2. Find all my-blog.json files without paired llms.txt
3. For each file (existing or new):
   • If new: Create complete structure from my-blog.json
   • If existing: Read last-updated date from metadata
   • Determine folder context (site-wide or folder-specific)
   • Fetch query-index.json from production
   • Filter by context (folder path matching)
   • Filter new posts (lastModified > last-updated)
   • Add new posts to organized sections (if any)
   • Update version date and metadata

/update-my-blog

Finds all my-blog.json files and updates them with latest content:

/update-my-blog

Process:

1. Find all my-blog.json files: **/*my-blog.json
2. For each file:
   • Read current content and last-updated
   • Determine folder context
   • Fetch query-index.json
   • Filter by context and date
   • Update categories and counts
   • Update metadata

Best Practices

Content Organization

✅ Use subsections - Break large categories into logical groups ✅ Consistent formatting - Use same pattern throughout ✅ Rich descriptions - Include full descriptions from source ✅ Clear hierarchy - Main section → Subsection → Posts ✅ Cross-references - Link between related llms.txt files

URL Management

✅ Source of truth: my-blog.json for URLs ✅ Text source: Existing llms.txt for descriptions/formatting ✅ New content: query-index.json for latest posts ✅ Relative URLs: Use relative paths in my-blog.json ✅ Absolute URLs: Use full URLs in llms.txt

Date Handling

✅ ISO format: YYYY-MM-DD for dates ✅ Comparison: Filter by lastModified > last-updated ✅ Display format: "Month Year" in llms.txt header ✅ Update always: Change date when content changes

Folder Context

✅ Use folder name: Final path segment indicates scope ✅ Filter appropriately: Only relevant content for folder llms.txt ✅ Paired files: Keep my-blog.json with llms.txt ✅ Independent updates: Each folder maintains own files

Examples

Example 1: Root llms.txt

Location: /llms.txt Scope: Entire site my-blog.json: /my-blog.json (all content) Categories: All site categories

Example 2: Integrations llms.txt

Location: /blogs/ddt/integrations/llms.txt Scope: EDS integration tutorials only my-blog.json: /blogs/ddt/integrations/my-blog.json Categories: Filtered to integration-related posts

Example 3: AI Topics llms.txt

Location: /blogs/ddt/ai/llms.txt Scope: AI/LLM content only my-blog.json: /blogs/ddt/ai/my-blog.json Categories: Filtered to AI-related posts

Troubleshooting

my-blog.json Missing

Solution: Create it from query-index.json filtered by folder context

Dates Not Comparing

Solution: Ensure ISO format (YYYY-MM-DD) and proper parsing

Wrong Content Scope

Solution: Check folder context and filter query-index.json properly

Duplicate Entries Within Categories

Solution: Deduplicate by URL before adding to categories

Related Files

• .claude/commands/update-llms.md - Slash command for updating llms.txt
• .claude/commands/update-my-blog.md - Slash command for updating my-blog.json
• scripts/sync-blog-content.js - Utility script for bulk updates (no dependencies)

Version

Skill Version: 1.0 Last Updated: November 2025 Line Count: < 500 ✅

---

Converted and distributed by TomeVault — claim your Tome and manage your conversions.