Skip to main content

tech-tutorial-writer

Generate high-quality, publish-ready technical tutorials for beginner, intermediate, and advanced readers. Use this skill whenever the user wants to write a tech article, tutorial, or blog post — especially for platforms like WeChat Official Account or Juejin. Trigger when the user says things like "write a tutorial", "help me write an article", "generate a blog post", "explain X in a structured way", or asks to produce educational content about a programming topic. Covers frontend, backend, algorithms, and AI/ML domains.

Datos de origen

Repositorio
OPBR/opbr-skills
Última actividad en el origen
10 de abril de 2026 a las 14:42
Idioma detectado de SKILL.md
inglés
Estrellas
1
Forks
0

Opciones de instalación

De forma predeterminada está seleccionado el prompt que primero revisa el origen. Puedes cambiar a un comando directo o descargar una copia local.

Revisa los archivos de origen

Lee SKILL.md y los archivos complementarios que muestra SkillsMP antes de decidir si quieres instalarlo.

Mostrando SKILL.md

SKILL.md
Instrucciones de origen · Vista previa de solo lectura
name
tech-tutorial-writer
description
Generate high-quality, publish-ready technical tutorials for beginner, intermediate, and advanced readers. Use this skill whenever the user wants to write a tech article, tutorial, or blog post — especially for platforms like WeChat Official Account or Juejin. Trigger when the user says things like "write a tutorial", "help me write an article", "generate a blog post", "explain X in a structured way", or asks to produce educational content about a programming topic. Covers frontend, backend, algorithms, and AI/ML domains.
# Tech Tutorial Writer — Skill Overview This skill is a structured writing system for producing **high-quality, publish-ready technical tutorials**. Every article it generates is designed to be dropped directly into a blog, developer community, or newsletter with minimal editing. --- ## Core Purpose The skill generates technical tutorials that guide readers through a **three-tier learning progression**: | Tier | Target Reader | Goal | |------|--------------|-------| | 🌱 **Beginner** | Zero experience / career changers | Build mental models, run the first working demo, eliminate fear | | 🚀 **Intermediate** | Has basics but hitting a ceiling | Understand the "why", master patterns, apply engineering best practices | | 🏆 **Advanced** | Wants to break to the next level | Source-code-level analysis, architecture thinking, production war stories | Before writing anything, the skill confirms which tier the article belongs to — because each tier has a completely different writing strategy. --- ## Article Structure (Required Sections) Every tutorial follows this skeleton: ### 1. Hook Opens with a relatable real-world pain point. The reader should think "that's exactly my problem." Then previews what they'll gain and how long it takes. ### 2. Mental Model Uses everyday analogies to explain abstract concepts before touching any code. Supports ASCII diagrams or Mermaid charts for visualization. No jargon without an explanation. ### 3. Environment Setup Lists exact dependency versions, provides a copy-paste initialization command, and pre-warns about common setup pitfalls. ### 4. Core Content Depth varies by tier: - **Beginner**: Every step has full runnable code + line-by-line breakdown + "what you see now" descriptions - **Intermediate**: Wrong approach shown first, then correct approach (contrast format), with supporting performance data - **Advanced**: Source code dives, architectural evolution (v1 → v2 → v3), real production incident examples ### 5. Hands-on Project Every article must include one complete, runnable demo. Code has inline comments on every key line. Project structure is shown explicitly. ### 6. FAQ 3–5 high-frequency questions in a consistent format with code examples in the answers. ### 7. Summary + Next Steps Bullet-point recap of what was learned, a clear "what to study next" path, and 2–3 curated resources. --- ## Code Quality Standards All generated code must: - Run without hidden dependencies - Include error handling (try/catch, error boundaries, etc.) - Follow the latest best practices for that language/framework - Clearly flag any deprecated APIs - Include TypeScript types for front-end tutorials --- ## Output Format Each generated article comes with a **metadata block** followed by the full article body: ``` ════════════════════════════ 📌 Article Metadata ════════════════════════════ Title (WeChat version): ... Title (Juejin version): ... Tier: Beginner / Intermediate / Advanced Summary (for Juejin description field): ... Recommended Tags: tag1, tag2, tag3 Estimated Word Count: ~XXXX words Target Platform: WeChat / Juejin / Both ════════════════════════════ 📄 Full Article (Markdown) ════════════════════════════ [complete tutorial body] ``` --- ## Platform Formatting Rules The skill adapts its output style depending on the target platform: ### WeChat Official Account - Title under 15 Chinese characters, must use a number or question to create curiosity - First 100 characters must hook the reader (WeChat collapses the preview) - Paragraphs capped at 5 lines, heavy use of short sentences - Ends with an engagement prompt ("What problems have you run into? Drop a comment!") ### Juejin (Developer Community) - Longer, keyword-rich titles for SEO - Uses Juejin-specific Markdown features: `:::tip`, `:::warning`, `:::danger` callout blocks - Comparison tables for alternative approaches - 3,000–8,000 words for intermediate/advanced articles --- ## Domains Covered The skill has pre-mapped learning progressions for: - **Frontend**: HTML/CSS basics → component patterns & state management → micro-frontends & compiler internals - **Backend**: REST APIs & auth → caching, queues, distributed systems → high-concurrency architecture & SRE - **Algorithms**: Arrays/stacks/queues → dynamic programming & graph algorithms → system design & large-scale data - **AI/ML**: Python basics & API calls → Prompt Engineering, fine-tuning, RAG → distributed training, inference optimization, MLOps --- ## What Makes This Different from Just "Write a Tutorial" The skill is modeled on patterns from the highest-performing tutorials on freeCodeCamp, roadmap.sh, Juejin, and top WeChat tech accounts. The key principles it enforces: | Principle | Description | |-----------|-------------| | **Scenario-first** | Every article starts with a problem, never a definition | | **Progressive complexity** | Simplest case first, complexity added incrementally | | **Immediate feedback loops** | Every step produces a visible result | | **Error-friendly** | Common mistakes are shown explicitly, not avoided | | **Moderate knowledge density** | Each article focuses on 1–3 core concepts, never a knowledge dump | | **Visual aids** | Flowcharts, comparison tables, syntax-highlighted code blocks | | **Conversational tone** | Addresses the reader as "you", like a senior dev explaining to a colleague |
Ver en GitHub