build-resume

What this skill does

# Resume Builder Skill

You are helping the user build and maintain their professional resume based on evidence from their actual work—not just memory.

---

## CRITICAL: Confidentiality Mandate

**Never extract or include company secrets, proprietary information, or sensitive data.**

When reviewing any source material:
- **DO** describe what the person accomplished and the impact they had
- **DO** identify skills, methodologies, and areas of expertise demonstrated
- **DO** note metrics if they're public or shareable (growth rates, scale, outcomes)
- **DON'T** copy proprietary content verbatim (code, strategies, processes)
- **DON'T** include internal project names, codenames, or client names without permission
- **DON'T** reveal trade secrets, confidential strategies, or unpublished work
- **DON'T** include credentials, internal URLs, or private contact information

The goal is to describe *what the person did* and *what skills they demonstrated*—not to document their employer's intellectual property. When in doubt, abstract it: "Led go-to-market strategy for enterprise product launch" not "Led the ProjectX launch targeting CompanyY's customers."

---

## Evidence Sources

This skill can help gather resume content from multiple sources:

| Source | What to Look For |
|--------|------------------|
| **Git Repositories** | Commits, tech stack, architecture patterns, project scope |
| **Documents** | Reports, proposals, strategies, presentations you authored |
| **Calendars** | Projects you led, meetings you ran, initiatives you drove |
| **Webpages** | Published articles, portfolio pieces, press mentions, bios |
| **LinkedIn** | Role history, recommendations, endorsements, posts |
| **Project Tools** | Jira/Asana tickets, completed initiatives, team contributions |
| **Published Work** | Papers, articles, designs, patents, open source contributions |
| **Performance Reviews** | Documented achievements, feedback, promotions |

---

## Master vs. Tailored Resume

This skill uses a two-tier approach:

- **Master Resume**: Your complete, unabridged work history. All roles, all projects, full detail. No length limit. This is your source of truth—maintain it over time, never send directly to employers.

- **Tailored Resume**: A 1-2 page version derived from the master, customized for a specific job posting. Lead with relevant experience, incorporate keywords, remove less relevant content. Generate fresh for each application.

---

## Available Modes

When invoked, ask which mode the user wants:

1. **Update Resume** - Add new roles, projects, or refine existing content in the master
2. **Gather Evidence** - Scan available sources for resume-worthy accomplishments
3. **Generate Tailored Version** - Create a targeted version from the master for a specific job
4. **Export to PDF/Word** - Convert a markdown resume to PDF or Word format
5. **Full Workflow** - Walk through the complete resume building process

---

## Mode: Update Resume

1. Ask where the user's current resume lives (or start fresh)
2. Ask what they want to add or update
3. For each new item, interview the user:
   - What was the **situation or challenge** you faced?
   - What **actions** did you take?
   - What were the **results or impact**?
   - What **skills** did this demonstrate?
   - Can you **quantify** the outcome?
4. Write content using the established format:
   - Situation → Action → Result framing
   - Concrete metrics when available
   - Active voice with strong verbs
   - Appropriate scope (team vs. individual contribution)

---

## Mode: Gather Evidence

1. Ask what sources are available to scan
2. For each source type:

**Git Repositories:**
- Find commits by author and date range
- Identify tech stacks from config files
- Note architectural patterns and project scope
- Suggest bullet points based on contribution volume

**Documents & Files:**
- Read reports, proposals, or presentations the user points to
- Extract key accomplishments and outcomes mentioned
- Identify themes and areas of expertise

**Webpages & Published Work:**
- Fetch articles, portfolio pages, or company announcements
- Pull quotes, metrics, or recognition mentioned
- Note the reach or impact of published work

**Calendar (if accessible):**
- Identify recurring meetings the user led
- Find project kickoffs and completions
- Note cross-functional collaboration patterns

3. Present findings organized by theme or time period with:
   - Source and date range
   - Key accomplishments discovered
   - Suggested resume bullet points
   - Questions to clarify with the user

---

## Mode: Generate Tailored Version

Create a resume tailored to a specific job posting.

### Input Options

The user can provide the job description via:
- **URL** — Fetch the job posting directly (e.g., LinkedIn, Greenhouse, Lever, company careers page)
- **Pasted text** — Job description copied into the conversation
- **File path** — Local file containing the job description

### Process

1. **Get the master resume**
   - Ask where the user's master resume lives
   - Read and parse the full content

2. **Get the job description**
   - If URL provided: fetch the page and extract the job posting
   - If text provided: use directly
   - If file provided: read the file

3. **Analyze the job requirements**
   - Extract required skills and technologies
   - Identify preferred/nice-to-have qualifications
   - Note years of experience requirements
   - Identify key responsibilities
   - Look for cultural or soft skill signals

4. **Map experience to requirements**
   - For each job requirement, find matching experience from the master resume
   - Identify gaps (requirements not covered)
   - Identify strengths (areas where experience exceeds requirements)
   - Note transferable skills that apply even if not exact matches

5. **Generate tailored resume**
   - Reorder sections to lead with most relevant experience
   - Emphasize bullet points that match job requirements
   - Incorporate keywords from the posting naturally (don't keyword stuff)
   - De-emphasize or remove less relevant experience
   - Adjust technical skills section to highlight relevant technologies
   - Keep the strongest metrics that demonstrate relevant impact
   - Fit to 1-2 pages

6. **Output**
   - Clean markdown ready for conversion to PDF
   - Summary of what was emphasized/de-emphasized
   - Notes on any gaps the user might want to address in a cover letter

### Example Interaction

```
User: /build-resume
Claude: Which mode? 1) Update 2) Gather Evidence 3) Tailored Version 4) Full Workflow
User: 3
Claude: Where is your master resume?
User: ./examples/master-resume.md
Claude: Got it. Now provide the job description - URL, paste the text, or give me a file path.
User: https://jobs.lever.co/company/senior-backend-engineer
Claude: [Fetches and analyzes the posting, generates tailored resume]
```

---

## Mode: Export to PDF/Word

Convert a markdown resume to a polished PDF or Word document.

### Prerequisites

Check if the user has the required tools installed:

```bash
# Check for pandoc (required)
pandoc --version

# For PDF output, one of these is needed:
# Option 1: wkhtmltopdf (recommended - simpler)
wkhtmltopdf --version

# Option 2: LaTeX (more control but heavier)
pdflatex --version
```

### Installation (if needed)

**macOS:**
```bash
brew install pandoc
brew install --cask wkhtmltopdf
```

**Ubuntu/Debian:**
```bash
sudo apt install pandoc wkhtmltopdf
```

**Windows:**
```bash
choco install pandoc wkhtmltopdf
```

### Export Process

1. **Get the resume file**
   - Ask for the path to the markdown resume
   - Verify the file exists and is valid markdown

2. **Ask for output format**
   - PDF (recommended for applications)
   - Word/DOCX (if the employer requires it)

3. **Ask for output location**
   - Default: same directory as source, with appropriate extension
   - Or user-specified path

4. **Generate the document**

**For PDF (using wkhtmltopdf):**
```bash
pando