Technical Blog Engine & Verification Pipeline

Overview

Generate production-ready, expert-level technical AI engineering articles with GEO/AEO direct answer blocks, factual checks, and build verification.

Verification & Test Notes

[!NOTE] Tested & Reviewed on 2026-10-10 for runtime environment Claude Code 2.x, Node.js 22, Astro 5.

Produces valid frontmatter, strict heading hierarchy, and zero hallucinated commands. All generated code blocks pass syntax validation.

Installation & Setup

Option 1: Direct Skill Download

1. Download Technical Blog Engine & Verification Pipeline (blog-generator.zip).
2. Unpack the zip file into your workspace agent directory:
   • Antigravity IDE: .agents/skills/blog-generator/
   • Claude Code: ~/.claude/skills/blog-generator/ or .agents/skills/blog-generator/
   • Cursor: .cursor/rules/blog-generator/ or your project prompt definitions.

Option 2: CLI Activation

If using Antigravity or compatible agent runners:

# Load into active workspace
mkdir -p .agents/skills/blog-generator
curl -sL https://nadhebe.com/skills/blog-generator.zip -o /tmp/blog-generator.zip
unzip /tmp/blog-generator.zip -d .agents/skills/blog-generator/

Skill Instructions & Reference

Nadhebe Master Technical Blog Generation & Publishing Engine

Use this skill whenever a user provides a topic, article idea, video transcript, or raw notes to generate production-ready, expert-level technical articles for Nadhebe.

1. End-to-End Blog Writing Workflow

Every article generated for Nadhebe must execute this sequential workflow:

┌─────────────────────────┐
│ 1. RESEARCH & INTENT    │ → Identify primary keyword, search intent, AEO queries,
│                         │   and analyze competing SERP pages (via `keyword-research` & `ai-seo`)
└───────────┬─────────────┘
            ▼
┌─────────────────────────┐
│ 2. FACT VERIFICATION    │ → Verify technical claims against official vendor docs.
│                         │   Strict zero-fabrication: NO fake stats, fake quotes, or anecdotes.
└───────────┬─────────────┘
            ▼
┌─────────────────────────┐
│ 3. COLLECTION SELECTION │ → Assign to most relevant collection: `guides`, `tutorials`,
│                         │   `comparisons`, `news`, `tool-reviews`, or `best-practices`.
└───────────┬─────────────┘
            ▼
┌─────────────────────────┐
│ 4. DRAFTING & GEO/AEO   │ → Write with single H1, 2-3 sentence direct answer blocks under H2s,
│                         │   structured comparison tables, code blocks, and Mermaid diagrams.
└───────────┬─────────────┘
            ▼
┌─────────────────────────┐
│ 5. HERO ART GENERATION  │ → Generate 16:9 Vintage Editorial Collage artwork via `generate_image`,
│                         │   convert to WebP and save to `public/images/<slug>-hero.webp`.
└───────────┬─────────────┘
            ▼
┌─────────────────────────┐
│ 6. PRE-PUBLISH AUDIT    │ → Run `content-quality-auditor` checklist, verify internal links,
│                         │   and ensure `npm run build` passes with 0 broken links.
└─────────────────────────┘

2. Content Collection Routing

Assign the article to the exact matching directory under src/content/:

• src/content/tutorials/: Hands-on step-by-step developer code walkthroughs, CLI setup guides, MCP tools, serverless deployments.
• src/content/guides/: Enterprise architecture breakdowns, pricing matrices, CI/CD pipeline automation, context engineering playbooks.
• src/content/comparisons/: Head-to-head model/tool benchmark reports (e.g. vLLM vs Ollama, Claude Code vs Cursor).
• src/content/news/: Major model releases, breaking benchmark announcements.
• src/content/tool-reviews/: In-depth single product/model evaluation reviews.
• src/content/best-practices/: Production engineering patterns, security audits, and latency optimization.

Do not force articles into irrelevant collections.

3. Article Markdown Frontmatter Schema

Every generated Markdown file must strictly conform to Nadhebe’s Astro content schema (src/content/config.ts):

---
title: "Primary Keyword First: Action / Benefit Target | Nadhebe"
description: "Comprehensive 150-160 character meta description containing primary keyword."
pubDate: 2026-10-10
updatedDate: 2026-10-10
author: nadhebe-team
category: "Guides" | "Tutorials" | "Comparisons" | "News" | "Reviews"
tags: ["Keyword1", "Keyword2", "Keyword3"]
heroImage: "/images/<slug>-hero.webp"
heroAlt: "Minimalist vintage editorial illustration representing <topic> on a <color> background"
estimatedReadingTime: 12
isPillar: true | false
topic: "Topic Cluster Name" # e.g. "AI Tool Costs", "Model Serving"
searchIntent: "informational" | "commercial" | "tutorial" | "comparison"
parentPillar: "<parent-slug>" # Optional (only if isPillar is false)
draft: false
itemsCompared: ["Item A", "Item B"] # Optional (only for comparisons)
faq:
  - question: "Direct user query?"
    answer: "Clear, factual, authoritative answer block."
  - question: "Another frequent question?"
    answer: "Concise answer without marketing fluff."
sources:
  - label: "Official Documentation"
    url: "https://example.com/docs"
  - label: "GitHub Repository / Benchmarks"
    url: "https://github.com/..."
---

4. Technical Depth, GEO/AEO & Editorial Standards

• Strict Factual Accuracy: Never fabricate statistics, token pricing, benchmarks, or citations. If a number is unknown or estimated, state it explicitly as directional.
• Zero AI Clichés: Prohibit empty filler words (“In today’s fast-paced digital world”, “delve into”, “game changer”, “seamlessly”).
• Direct Answer Blocks (GEO/AEO): Under every H2, write a self-contained 2–3 sentence direct answer block that can be cited in Google AI Overviews or Perplexity.
• Structured Data Tables: Format complex pricing, feature matrices, or benchmark results into clean Markdown tables.
• Originality Threshold: Minimum 70% original synthesis, code examples, Mermaid diagrams, performance tables, and decision trees.

5. Hero Artwork Specification (Vintage Editorial Collage)

All hero images generated for Nadhebe blogs must follow the Vintage Editorial Mixed-Media Collage style:

Aesthetic Principles:

• Canvas / Background: Flat pastel or muted neutral tones (Sage Green #87A987, Soft Olive #A3B18A, Warm Sand #F4F1DE, Soft Terracotta #E07A5F, Muted Mint #B7E4C7, Soft Lavender #D8B4F8).
• Visual Subject Cutouts: High-contrast black-and-white or duotone cutouts of tech subjects (terminals, server racks, network graph nodes, calculators, keyboards) paired with subtle botanical leaves.
• Geometric Elements: Concentric circles, solid yellow/orange accent suns (#F4A261), and vintage halftone textures.
• Anti-Patterns: ❌ ZERO neon 3D glowing spheres, ❌ ZERO muddy black backgrounds, ❌ ZERO artificial text watermarks.

Image Prompt Template for generate_image:

Vintage editorial collage illustration on a [COLOR PALETTE] background. A black and white cutout of [SUBJECT A] and [SUBJECT B], surrounded by warm yellow geometric circles, vintage magazine halftone textures, and subtle green plant leaves. Clean minimal tech art style, 16:9 ratio

6. Post-Generation Execution & Build Verification

After creating or updating blog posts:

1. Convert the generated artifact image to WebP and save to public/images/<slug>-hero.webp.
2. Run npm run build to compile static HTML pages and trigger Pagefind search indexing.
3. Verify that scripts/verify-links.ts passes with 0 broken links.
4. Run git add ., git commit -m "...", and git push to deploy to GitHub.