Tag Article Skill

Analyze archive article content and apply relevant tags from the canonical tag list.

Usage

/tag-article <filename>

Example:

/tag-article 2020-03-12-git-is-for-everyone.md

Workflow

1. •Read the article from public/archive/<filename>
2. •Load canonical tags from public/archive/tags.json
3. •Analyze content for tag indicators:
   • •Title keywords and patterns
   • •Code block languages (js, csharp, etc.)
   • •Technical terms and framework mentions
   • •Content depth and type
4. •Select 3-6 appropriate tags based on analysis
5. •Update manifest.json with the tags array
6. •Report what tags were applied

Tag Selection Guidelines

Analyze These Content Signals

1. •Title keywords: Map words in the title to tags using the keywords mapping in tags.json
2. •Code blocks: Identify language hints from fenced code blocks
3. •Technical terms: Look for framework names, libraries, and domain-specific vocabulary
4. •Content type patterns:
   • •"Quick Post" → quick-tips
   • •"Book Review" → reviews, learning
   • •Tutorial-style content → teaching
   • •Personal experiments → experiments
   • •Philosophy/reflection → philosophy
   • •Historical content → history

Tag Quantity Rules

• •Minimum: 2 tags (content_type + at least one technical tag)
• •Maximum: 6 tags (avoid over-tagging)
• •Ideal: 3-5 tags

Tag Priority

1. •Required: Always include one content_type tag
2. •Primary: Main technical domain (graphics, games, web, etc.)
3. •Secondary: Specific technologies mentioned (unity, react, etc.)
4. •Languages: Only if code examples are shown or language is discussed

Examples

Article about Unity physics:

["unity", "games", "physics", "c-sharp", "teaching"]

Philosophical post about software:

["philosophy", "software"]

Quick tip about JavaScript:

["quick-tips", "js", "web"]

WebGL graphics experiment:

["experiments", "webgl", "graphics", "js"]

Implementation

When the skill is invoked:

1. Parse the filename from arguments
2. Read public/archive/<filename>
3. Read public/archive/tags.json
4. Read public/archive/manifest.json

5. Analyze article content:
   - Extract title from frontmatter
   - Find code blocks and their language hints
   - Scan for keywords from tags.json keywords mapping
   - Determine content type from title patterns and content

6. Select appropriate tags (3-6 tags)

7. Update manifest.json:
   - Find the post entry matching the filename
   - Add "tags" array to that entry
   - Write updated manifest.json

8. Report: "Applied tags [tag1, tag2, ...] to <filename>"

Data Structures

tags.json Structure

{
  "tags": [
    "ai", "akka", "aws", "azure", "backend", "bootstrap", "c", "cg", "c-sharp",
    "deep-dives", "docker", "dotnet", "experiments", "flash", "frontend",
    "games", "glsl", "go", "graphics", "haxe", "history", "java", "js",
    "learning", "liveops", "mobile", "nakama", "nodejs", "objc", "physics",
    "python", "react", "redux", "software", "solidity", "teaching", "ts",
    "typescript", "unity", "web", "web3", "webgl"
    // ... full list in tags.json
  ],
  "keywords": {
    "Unity": ["unity", "games", "c-sharp"],
    "shader": ["graphics", "cg", "glsl"],
    ...
  }
}

manifest.json Post Entry (after tagging)

{
  "title": "Example Post",
  "date": "2020-01-01T00:00:00.000Z",
  "filename": "2020-01-01-example-post.md",
  "originalUrl": "https://...",
  "tags": ["teaching", "unity", "games", "c-sharp"]
}

Notes

• •Tags are applied without confirmation (auto-apply)
• •The manifest.json is updated in place
• •If the article already has tags, they will be replaced
• •Use the keywords mapping as a guide, but also apply judgment for unlisted terms