Content

Intent Routing

Content Creation Intents

Trigger patterns:

• "Create/write/draft new [blog post|article|whitepaper|case study]"
• "Start a post about [topic]"
• "Initialize content for [subject]"

Routes to: InitContent workflow → init-content.ts tool

What happens:

1. Prompt for content type (blog, whitepaper, case study)
2. Gather metadata (title, category, tags, target audience)
3. Render appropriate template with frontmatter
4. Save to _drafts/ directory
5. Open in editor for writing

Content Review Intents

Trigger patterns:

• "Review this [post|article|whitepaper]"
• "Check this content for [SEO|accuracy|brand voice]"
• "Analyze this draft"
• "Get feedback on this article"

Routes to: ReviewContent workflow → review-content.ts tool

What happens:

1. Auto-detect content file (or accept path parameter)
2. Extract frontmatter and content
3. AI analysis for:
   • Technical accuracy
   • Brand voice alignment (against Brand-Guidelines.md)
   • SEO optimization (title, meta, keywords, readability)
   • Grammar and style
   • Example validation (if code/technical)
4. Generate review report with specific recommendations
5. Optionally apply automated fixes

Content Publishing Intents

Trigger patterns:

• "Publish this draft"
• "Move [filename] to published"
• "Categorize and publish this content"
• "Ready to publish [topic] post"

Routes to: PublishContent workflow → publish-content.ts tool

What happens:

1. Find draft file
2. Validate frontmatter completeness
3. Determine target path based on type/year/category
4. Move from _drafts/ to appropriate published folder
5. Update content index in Content/README.md
6. Update status to "Published" in frontmatter
7. Backup before any modifications

Content Distribution Intents

Trigger patterns:

• "Generate social snippets for [post]"
• "Create LinkedIn post for this article"
• "Format this for [platform]"
• "Create distribution package"

Routes to: DistributeContent workflow → distribute-content.ts tool

CreateContentFromUpgrades Intents

Trigger patterns:

• "Create content from [Anthropic updates|YouTube video|upgrades]"
• "Turn [these updates|this video] into [blog post|article|content]"
• "Make content from [upgrade information]"
• "Generate article from [release notes|update]"

Routes to: CreateContentFromUpgrades workflow → transform-to-content.ts tool

What happens:

1. Identify input source (Anthropic updates or YouTube video)
2. Get upgrade data (from Upgrades skill or direct input)
3. Transform upgrade data to structured content
4. Generate content draft with SEO optimization
5. Save to _drafts/ directory with proper frontmatter
6. Ready for review and publishing

What happens:

1. Load published content
2. Generate platform-specific snippets:
   • LinkedIn (1-2 paragraphs + link)
   • Twitter/X (thread format)
   • Dev.to/Hashnode (full formatted article)
   • Email newsletter (summary format)
3. Extract key quotes for social media
4. Create visual content suggestions
5. Save distribution package to [content]/_distribution/

Workflow Architecture

User Intent → Skill Router → Workflow → CLI Tool → Processing → Output

Example: "Create new blog post about PAI"

Layer 1: INTENT ROUTING
  "Create new blog post about PAI"
  ↓ Matches "Create/write/draft new [blog post]"
  ↓ Routes to InitContent workflow

Layer 2: WORKFLOW ORCHESTRATION
  InitContent.md determines:
  - Prompt for metadata (title, category, tags)
  - Select blog post template
  - Invoke init-content.ts tool

Layer 3: CLI TOOL EXECUTION
  init-content.ts runs with:
  - type: blog
  - title: "Building Personal AI Infrastructure"
  - category: "AI Infrastructure"
  - Auto-detects Content directory root

Layer 4: PROCESSING
  - template-renderer.ts renders blog-post.md.hbs
  - file-ops.ts writes to _drafts/ safely
  - content-index-manager.ts tracks new draft

Layer 5: OUTPUT
  - New file: _drafts/building-personal-ai-infrastructure-draft.md
  - Pre-filled frontmatter with metadata
  - Template sections ready to fill
  - Status: Draft

Content Root Detection

All tools auto-detect the Content directory by:

1. Looking for /Documents/Projects/Content/ directory
2. Reading Content/README.md for configuration
3. Falling back to user-specified path if needed

No manual path specification required.

Integration with Existing Systems

Brand Guidelines:

• Review workflow reads ~/Documents/Projects/Consulting/Business/Brand-Guidelines.md
• Validates brand voice, tone, color usage, design patterns

Client Projects:

• Case study workflow can link to client folders in ~/Documents/Projects/Consulting/Clients/
• Auto-extract project metadata from README.md

PAI Ecosystem:

• Uses shared libraries (file-ops, ai-processor from other packs)
• Composable with ClientDelivery (generate case studies from projects)

Content Types Supported

Blog Posts

• Thought leadership articles
• Technical tutorials
• Case study narratives
• Industry insights

Whitepapers

• In-depth technical guides
• Architecture deep-dives
• Research reports
• Best practices documentation

Case Studies

• Client success stories
• Project showcases
• Before/after analysis
• ROI demonstrations

Usage Examples

Create New Blog Post

User: "Create a new blog post about RAG architecture"

Ava: I'll create a new blog post draft about RAG architecture.
[Runs init-content.ts]

Created: _drafts/rag-architecture-best-practices-draft.md
- Type: Blog Post
- Category: Technical Architecture
- Status: Draft
- Word count target: 2000-3000

Ready for you to write. Open in your editor?

Review Content

User: "Review the PAI infrastructure blog post"

Ava: I'll analyze the PAI infrastructure post for technical accuracy, SEO, and brand voice.
[Runs review-content.ts]

Review complete:

Technical Accuracy: ✅ Excellent
- All architecture claims validated
- Code examples tested and working
- Technical terminology used correctly

SEO Optimization: ⚠️ Needs improvement
- Title: Good (60 chars)
- Meta description: Missing - add 150-160 char summary
- Keyword density: "PAI" appears 47 times (good)
- Readability: Grade 12 (consider simplifying some sections)
- Headings: Well-structured with H2/H3 hierarchy

Brand Voice: ✅ Aligned
- Professional but conversational ✓
- First-person perspective ✓
- No unnecessary emojis ✓
- Matches brand guidelines

Recommendations:
1. Add meta description for SEO
2. Simplify the "5-Layer Architecture" section intro
3. Add alt text suggestions for diagrams

Apply automated fixes? (Y/n)

Publish Content

User: "Publish the PAI blog post"

Ava: I'll publish the PAI infrastructure post.
[Runs publish-content.ts]

Validated frontmatter: ✓
Moving: _drafts/pai-infrastructure-draft.md
     → Blog/2026/AI/pai-personal-ai-infrastructure.md

Updated Content/README.md index:
| From AI Assistant to AI Infrastructure | Blog | AI | Published | Blog/2026/AI/pai-personal-ai-infrastructure.md |

Status: Published ✓

Next steps:
1. Run "Generate social snippets for PAI post"
2. Publish to pragsys.io/blog
3. Share on LinkedIn

Generate Distribution Package

User: "Generate social snippets for the PAI post"

Ava: I'll create a distribution package for the PAI infrastructure post.
[Runs distribute-content.ts]

Created distribution package at:
Blog/2026/AI/_distribution/pai-personal-ai-infrastructure/

Files generated:
- linkedin-post.md (2 paragraphs + link)
- twitter-thread.md (8-tweet thread)
- dev-to-article.md (full formatted for Dev.to)
- email-newsletter.md (summary format)
- key-quotes.md (5 shareable quotes)
- visual-suggestions.md (diagram/image ideas)

LinkedIn preview:
---
I spent months using AI as an assistant. Then I realized: I didn't need an AI assistant—I needed AI infrastructure.

PAI (Personal AI Infrastructure) transforms Claude Code from a helpful chat partner into a complete automation platform. Real example: 29-minute discovery call → 3,085 lines of professional documentation, automatically.

Read about the architecture, real-world results, and why "infrastructure thinking" beats "task thinking": [link]
---

Ready to distribute?

Advanced Features

Content Analytics (Future)

• Track views, engagement, conversions
• A/B test different titles/intros
• Suggest optimizations based on performance

Multi-Author Support (Future)

• Author profiles and attribution
• Collaborative review workflows
• Version control integration

Content Calendar (Future)

• Schedule publishing dates
• Track editorial pipeline
• Remind about stale drafts

Technical Stack

• Runtime: Bun
• Language: TypeScript
• Templates: Handlebars
• AI Processing: Anthropic Claude API
• SEO Analysis: Custom algorithms + AI validation

Installation

See docs/INSTALL.md for full installation instructions.

Quick start:

cd ~/.claude/skills/Content
bun install
export ANTHROPIC_API_KEY="your-key"
bun run src/tools/init-content.ts --help

Status

Version: 1.0.0 (In Development) Completion: 0% (Just created) Next: Implement workflows and tools

Related Skills

• CORE - Personal identity and preferences (auto-loaded)
• ClientDelivery - Consulting project automation (compose for case studies)
• Art - Visual content generation (integrate for diagrams)

This skill is part of the PAI (Personal AI Infrastructure) ecosystem.