Format Cover Letter

How It Works

1. Semantic Analysis: I analyze content to understand elements (contact info, recipient address, RE line, body text, etc.)

2. Style Mapping: Based on context, I assign appropriate styles:

   • "ANTHONY BYRNES" at top → Contact Name
   • "RE: Position Title" → RE Line (bold orange)
   • Body text mentioning "Louis & Keely: Live at the Sahara" → Auto-italicized via dictionary

3. Hybrid Inline Styling: Play/production titles automatically italicized:

   • Dictionary lookup (95% of cases - no manual markup needed)
   • Manual override for new/unknown plays
   • Exclusion for edge cases

4. Generate Document: Create formatted .docx using template

5. Visual Preview: Show PDF preview for review

6. Learn from Corrections: If you correct a style choice, I remember for next time

The 19 Styles

Paragraph Styles:

• CV Name - Your name at top
• Contact Name - Bold, 10pt (cover letter header)
• Contact Info - Regular, 10pt (phone/email)
• Recipient Address - Regular, 11pt (organization address)
• RE Line - Bold, 13pt, ORANGE (position/subject line)
• Date Line - Right-aligned, 11pt (document date)
• Section Header - Bold, ORANGE (11pt for CV, 13pt for cover letter)
• Body Text - Standard paragraphs, 11pt
• Timeline Entry - Date + institution with hanging indent (CV only)
• Page Header - Bold, 10pt (page 2+ headers)
• Bullet Standard, Bullet Gray, Bullet Emphasis - List variations

Character Styles (inline):

• Play Title - Italic for productions (auto-styled via dictionary!)
• Institution - Bold for school/company names
• Job Title - Bold italic for positions
• Orange Emphasis - Highlight text
• Gray Text - Dates, secondary info

Cover Letter Elements

Contact Block (top of page):

ANTHONY BYRNES
T: 213.305.3132
E: [email protected]

Recipient Address:

Colburn School
200 South Grand Avenue
Los Angeles, CA 90012

RE Line (bold orange, same as section headers):

RE: General Manager of Performances & Events Division

Body Text with auto-italicized play titles:

For Louis & Keely: Live at the Sahara, I generated $1.4 million...

→ "Louis & Keely: Live at the Sahara" automatically italicized from dictionary

Signature:

Sincerely,

[signature image]

Anthony Byrnes

Metadata Inference (NEW!)

You provide: Raw text or minimal JSON (just body paragraphs)

I infer:

• Contact block (from defaults.yaml)
• Date line (today's date)
• Recipient address (from job description or content)
• RE line (from job title)
• Salutation (extracted from content)

Smart Job Description Matching:

1. I search career-applications/*/00-job-description.md files
2. Match based on organization/position mentioned in your letter
3. Confirm with you: "Found job description for UCLA - Associate Dean. Use this?"
4. Extract recipient and RE line from confirmed job description

Confirmation Before Formatting:

Inferred Metadata:

Contact:       ✓ ANTHONY BYRNES
               ✓ T: 213.305.3132
               ✓ E: [email protected]

Date:          November 11, 2025

Recipient:     UCLA School of Theater, Film and Television

RE:            Associate Dean and Chief Administrative Officer [CAO]

Salutation:    Dear Search Committee,

Looks correct? (yes/modify/[field name])

Configuration:

• Defaults: ~/.claude/skills/format-cover-letter/defaults.yaml
• Job descriptions: Auto-searched in career-applications/

Workflow

Step 1: Analyze Content

I'll parse your content and create JSON structure:

{
  "document_metadata": {
    "type": "cover-letter",
    "author_name": "Anthony Byrnes",
    "document_title": "Position Title Cover Letter"
  },
  "content": [
    {"text": "ANTHONY BYRNES", "style": "Contact Name", "type": "paragraph"},
    {"text": "RE: Position Title", "style": "RE Line", "type": "paragraph"},
    {"text": "Body with play title...", "style": "Body Text", "type": "paragraph"}
  ]
}

Step 2: Generate Document

I'll call the formatter to create your .docx:

import json
import subprocess
from pathlib import Path

# Save content mapping
mapping_file = "/tmp/cover_letter_mapping.json"
with open(mapping_file, 'w') as f:
    json.dump(data, f, indent=2)

# Format document
result = subprocess.run([
    "python3",
    "format_cv.py",
    mapping_file,
    output_path,
    "--document-type", "cover-letter",
    "--preview"
], capture_output=True, text=True, cwd=str(Path.home() / "PycharmProjects/career-lexicon-builder"))

Step 3: Visual Preview

I'll convert to PDF and show you what it looks like using the open command.

Step 4: Review

You can request changes: "That venue should be italic" or "Don't emphasize that number"

Step 5: Learn (if corrections made)

If you made corrections, I'll update learned-preferences.yaml and add new plays to play-titles-dictionary.yaml.

Hybrid Inline Styling

The Problem: Manually marking up every play title is tedious:

{
  "text": "For Louis & Keely: Live at the Sahara, I...",
  "runs": [
    {"text": "For ", "style": null},
    {"text": "Louis & Keely: Live at the Sahara", "style": "Play Title"},
    {"text": ", I...", "style": null}
  ]
}

The Solution: Dictionary auto-styling (95% of cases):

{
  "text": "For Louis & Keely: Live at the Sahara, I...",
  "style": "Body Text"
}

→ "Louis & Keely: Live at the Sahara" automatically italicized from dictionary!

Dictionary Location:

~/.claude/skills/format-cover-letter/play-titles-dictionary.yaml

Example dictionary: