init-rules

Your Task

You are the init-rules skill for the hustler-rails plugin. When invoked, copy Rails coding conventions from the plugin to the user's project, filtering by detected dependencies.

Step 1: Detect Project Dependencies

Build comprehensive dependency list:

# Gems from Gemfile.lock
grep "^    " Gemfile.lock 2>/dev/null | awk '{print $1}' | sort

# Database adapter
grep "adapter:" config/database.yml 2>/dev/null | head -1 | awk '{print $2}'

# JS/UI from importmap
grep 'pin "' config/importmap.rb 2>/dev/null | awk -F'"' '{print $2}'

# JS/UI from CDN links in views
grep -rh "cdn\|unpkg\|jsdelivr" app/views/ 2>/dev/null | grep -oE "(beercss|tailwind|bootstrap|alpinejs|htmx)" | sort -u

# JS/UI from vendor assets
find vendor/javascript vendor/assets app/assets -type f 2>/dev/null | grep -oE "(beercss|turbo|stimulus|tailwind|bootstrap)" | sort -u

# JS/UI from package.json
grep -E '"(dependencies|devDependencies)"' package.json 2>/dev/null | grep -oE "(beercss|turbo|stimulus)" | sort -u

Normalize detected dependencies to lowercase. Build a single master list of all detected dependencies.

Step 2: Check for Existing Rules

Check if .claude/rules/hustler-rails/ exists and has files.

If existing rules found, use AskUserQuestion tool:

{
  "questions": [
    {
      "question": "hustler-rails rules already exist. How should I proceed?",
      "header": "Existing rules",
      "multiSelect": false,
      "options": [
        {
          "label": "Backup and recreate",
          "description": "Renames existing to .backup-[timestamp]"
        },
        {
          "label": "Merge new files",
          "description": "Keep existing, add new only"
        },
        { "label": "Cancel", "description": "Exit without changes" }
      ]
    }
  ]
}

Handle response:

• Backup and recreate: mv .claude/rules/hustler-rails .claude/rules/hustler-rails.backup-$(date +%Y%m%d-%H%M%S)
• Merge new files: Skip files that already exist
• Cancel: Exit without changes

Step 3: Find All Rule Files

Get list of all rule markdown files:

find {baseDir}/assets/rules -name "*.md" -type f | grep -v README.md

Step 4: Process Each Rule File

For each rule file found:

1. Read the file using Read tool
2. Parse front matter - Extract dependencies: [...] and examples: [...] arrays
3. Check dependencies:
   • If dependencies: [] (empty) → Rule is universal, COPY
   • If all dependencies in list are detected → COPY
   • If any dependency missing → SKIP
4. Copy rule file if dependencies satisfied:

   mkdir -p "$(dirname .claude/rules/hustler-rails/[relative-path])"
   cp [source-path] .claude/rules/hustler-rails/[relative-path]
   

5. Copy examples if front matter has examples: [...]:
   • For each example name in array:

     mkdir -p .claude/rules/hustler-rails/views/examples
     cp -r {baseDir}/assets/rules/views/examples/[name] .claude/rules/hustler-rails/views/examples/
     

Dependency matching notes:

• Match case-insensitively
• Handle database adapter aliases:
  • sqlite or sqlite3 → matches sqlite3 adapter
  • pg or postgresql → matches postgresql adapter
  • mysql or mysql2 → matches mysql2 adapter

Track:

• Rules copied count
• Rules skipped count
• Examples copied count
• List of matched dependencies
• List of skipped dependencies with reasons

Step 5: Report Results

Report detailed summary:

✅ hustler-rails rules initialized

📊 Results:
   Rules: [X] copied, [Y] skipped
   Examples: [Z] directories copied

📍 Destination: .claude/rules/hustler-rails/

Matched dependencies:
- pundit (Gemfile.lock)
- beercss (CDN in app/views/layouts/application.html.erb)
- sqlite3 (database.yml)

Skipped rules (missing dependencies):
- views/topics/tailwind/tailwind-components.md (requires: tailwind)
- authorization/topics/cancancan/cancancan-abilities.md (requires: cancancan)

List each copied example with the rule that referenced it.