Frontend Resume

Create zero-dependency, animation-rich HTML resume pages that run entirely in the browser. Help users discover their preferred aesthetic through visual exploration, then generate production-quality interactive resumes.

Core Philosophy

1. Zero Dependencies — Single HTML files with inline CSS/JS. No npm, no build tools.
2. Show, Don't Tell — Generate visual previews, let users pick what they like.
3. Distinctive Design — Avoid generic "AI slop" aesthetics. Each resume should feel custom-crafted.
4. Print-Friendly — Every resume MUST include @media print styles for clean PDF output.
5. Accessible — Semantic HTML, keyboard nav, reduced motion support, WCAG AA contrast.

Phase 0: Detect Mode

Determine what the user wants:

Mode A: From File — User provides a resume file (.pdf, .docx, .md, .txt, .json) → Phase 1 Mode B: From Text — User provides resume content as text → Phase 1 Mode C: Enhancement — User has an existing HTML resume to improve → Read file, then Phase 2

Phase 1: Content Extraction & Structuring

Step 1.1: File Extraction (Mode A only)

Run scripts/extract_resume.py on the user's resume file:

python [skill-dir]/scripts/extract_resume.py <resume_file>

Supported formats and dependencies:

If a required dependency is missing, the script prints install instructions.

Present extracted content to user for confirmation. For JSON Resume files, the data is already structured — present the mapped sections directly and skip to Step 1.3 unless the user wants to adjust.

Step 1.2: Structure Content

Parse raw text into structured sections:

• header: name, title/tagline, location, contact (email, phone, LinkedIn, GitHub, portfolio)
• about: professional summary
• experience[]: company, role, dates, location, bullets
• education[]: school, degree, dates, honors/GPA
• skills{}: grouped by category with optional proficiency levels (0-100)
• projects[]: name, description, tech stack, link
• certifications[]: name, issuer, date (optional)
• languages[]: language, proficiency (optional)

Present structured content to user for review. Ask: "Does this look correct? Any sections to add, remove, or modify?"

Step 1.3: Section Priority

Ask via AskUserQuestion:

• Header: "Layout"
• Question: "Which sections should be most prominent?"
• Options:
  • "Experience-heavy" — Work history is the star (senior professionals)
  • "Skills-focused" — Technical skills front and center (developers/engineers)
  • "Projects-driven" — Portfolio/projects spotlight (designers/freelancers)
  • "Balanced" — Equal weight to all sections

Phase 2: Style Discovery

Step 2.0: Style Path

Ask: "How would you like to choose your resume style?"

• "Show me options" — Generate 3 visual previews (recommended) → Step 2.1
• "I know what I want" — Pick from preset list directly → Show preset picker below

Available Presets:

If user picks one directly, skip to Phase 3.

Step 2.1: Guided Discovery

Question 1 — Mood (multiSelect up to 2):

• "Professional/Authoritative" — Established, trustworthy
• "Creative/Distinctive" — Unique, design-forward
• "Modern/Technical" — Clean, developer-oriented
• "Warm/Approachable" — Friendly, personable

Question 2 — Tone:

• "Light" — Clean, printable
• "Dark" — Bold, screen-optimized
• "Either" — Let the style decide

Step 2.2: Generate 3 Previews

Based on mood + tone, select 3 presets. Read references/STYLE_PRESETS.md for full preset definitions.

Mood-to-preset mapping:

Generate 3 mini HTML previews in .claude-design/resume-previews/:

.claude-design/resume-previews/
├── style-a.html
├── style-b.html
└── style-c.html

Each preview: self-contained (~80-120 lines), shows hero section + one experience entry + skill bars. Enough to convey the aesthetic.

NEVER use these generic patterns:

• Inter, Roboto, Arial as display fonts
• Purple gradients on white, generic indigo #6366f1
• Everything centered with no variation
• Gratuitous glassmorphism

Step 2.3: Choose

Present previews, ask user to pick one or mix elements.

Phase 3: Generate Resume

Read these reference files for the chosen preset:

• references/STYLE_PRESETS.md — Full CSS variables, font URLs, color palette, animation style
• references/RESUME_COMPONENTS.md — Interactive component CSS/JS patterns
• references/HTML_ARCHITECTURE.md — Base HTML template, mandatory CSS, JS architecture, print styles, accessibility checklist

Output

Generate a single self-contained HTML file:

• All CSS inline in <style>
• All JS inline in <script>
• Google Fonts / Fontshare via <link> (gracefully degradable)
• Semantic HTML5 structure

Required Features

1. Scroll-triggered reveals — Elements animate in via IntersectionObserver
2. Floating navigation — Fixed nav with active section tracking
3. Animated skill bars — Fill on scroll-into-view
4. Scroll progress bar — Thin bar at top
5. Back-to-top button — Appears after scrolling
6. Print stylesheet — @media print with white bg, no animations, visible URLs
7. Reduced motion — @media (prefers-reduced-motion: reduce) disables animations
8. Responsive — All sizes use clamp(), works mobile through desktop

Verification Checklist

Before delivering, verify:

• All typography uses clamp(min, preferred, max)
• All spacing uses clamp() or viewport units
• Semantic HTML (<header>, <main>, <section>, <nav>)
• Print styles included and functional
• Reduced motion respected
• ARIA labels on nav, buttons, progress bar
• Color contrast meets WCAG AA
• External links have target="_blank" rel="noopener"

Phase 4: Delivery

1. Clean up — Delete .claude-design/resume-previews/ if it exists
2. Open — Launch in browser: open [filename].html
3. Summary:

Your interactive resume is ready!

File: [filename].html
Style: [Style Name]
Sections: [list]

Features:
- Scroll-triggered animations
- Animated skill bars
- Interactive timeline
- Floating navigation
- Print-friendly (Ctrl+P)
- Responsive on all screen sizes

To customize:
- Colors: Edit :root CSS variables at the top
- Fonts: Change the font import link
- Animations: Modify .reveal class timings
- Content: Edit text directly in the HTML

Would you like me to make any adjustments?

Troubleshooting

PDF extraction returns little text: PDF may be image-based (scanned). Ask user to provide text content directly.

DOCX extraction missing structure: Some DOCX files don't use standard heading styles (Heading 1/2/3). The extractor falls back to font size from individual runs. If structure is still unclear, ask the user to confirm which lines are section headers.

JSON Resume fields missing: The JSON may omit optional sections (projects, certificates, etc.). Present what's available and ask the user if they want to add missing sections manually.

Markdown frontmatter interference: YAML frontmatter delimiters (---) are skipped automatically. If the resume content itself uses --- as horizontal rules, they will also be skipped — this is generally harmless.

Fonts not loading: Check font URL. Ensure font-family declaration includes system font fallback.

Animations not triggering: Verify IntersectionObserver is initialized. Check that .visible class is being added.

Print output looks wrong: Ensure @media print block exists. Check that animations are disabled and backgrounds are white.