blog-kit

Blog Kit

AI-powered blog article generation with specialized Claude Code agents.

Installation

Quick Install

# 1. Add plugin to marketplace
/plugin marketplace add https://github.com/leobrival/blog-kit.git

# 2. Install the plugin
/plugin install blog-kit

Detailed Installation Guide

For complete installation instructions, troubleshooting, and verification steps, see INSTALL.md.

What's included:

• ✅ 12 slash commands (/blog-*)
• ✅ 9 specialized agents
• ✅ JSON template system (3 templates, 8 components)
• ✅ Persona system (Jobs-to-be-Done, Forces of Progress)
• ✅ Complete documentation

Quick Start

# 1. Setup or analyze existing content
/blog-setup                    # New blog (manual setup)
/blog-analyse                  # Existing blog (auto-detect)

# 2. Generate your first article
/blog-generate "Your article topic"

Usage Examples

Per-Article Commands (Recommended)

Work on a specific article by providing the slug:

# Single article optimization
/blog-optimize "en/nodejs-logging"

# Single article translation
/blog-translate "en/nodejs-logging" "fr"

# Single article image optimization
/blog-optimize-images "en/nodejs-logging"

Token usage: ~5k-15k tokens per article

Global Commands (⚠️ High Token Usage)

Apply command to ALL articles in your blog:

# ⚠️  WARNING: This will analyze ALL articles (can use 50k-500k tokens)
/blog-optimize

# ⚠️  WARNING: This will validate translation coverage for ALL articles
/blog-translate

# ⚠️  WARNING: This will optimize images in ALL articles
/blog-optimize-images

Token usage: 50k-500k tokens depending on article count Cost: Can be expensive if you have many articles Use case: Initial setup, bulk operations, CI/CD pipelines

Batch Commands (Controlled)

Process specific articles efficiently:

# Optimize multiple specific articles
/blog-optimize "en/article-1"
/blog-optimize "en/article-2"
/blog-optimize "fr/article-3"

# Translate specific articles to multiple languages
/blog-translate "en/microservices" "fr"
/blog-translate "en/microservices" "es"
/blog-translate "en/microservices" "de"

Recommendation: Process articles individually or in small batches to control token usage and costs.

Features

• 100% Action-Oriented Agents: Every agent produces actionable content (drafts, briefs, or final articles)
• Behavioral Personas: Jobs-to-be-Done, Forces of Progress, Customer Awareness frameworks for audience targeting
• Multi-agent workflow: Research (→ draft) → SEO/GEO (→ briefs) → Marketing (→ article) → Quality (→ auto-fix)
• Automatic Draft Generation: Research agent now generates article drafts, not just reports
• Auto-Fix Quality Issues: Quality optimizer automatically corrects detected problems
• Batch Content Updates: Analyzer updates all existing articles with missing metadata
• JSON-based templates: ShadCN-inspired template system with strict validation
• Template library: Tutorial, Guide, Comparison templates with GEO components
• Hierarchical configuration: 5-level cascade (global → template → language → category → article)
• Context optimization: 99.5% token efficiency
• File-based handoffs: Zero context pollution
• User checkpoints: Review at each phase
• Pure plugin: No dependencies, scripts are local utilities only
• Internationalization: Multi-language article structure with automated translation
• Translation validation: Automated i18n coverage reports and missing translation detection
• Image optimization: Automated WebP conversion (80% quality)
• GEO optimization: AI search optimization (ChatGPT, Perplexity, Google AI Overviews)
• Post Types & Funnel Stages: Content adapts to 4 post types (actionnable/aspirationnel/analytique/anthropologique) and 3 funnel stages (TOFU/MOFU/BOFU)

Commands