job-apply

What this skill does

# Job Application Assistant

A Claude Code skill for filling job applications on LinkedIn Easy Apply, Greenhouse, Ashby, Lever, Rippling, and Workday using browser automation.

## Initial Prompt

When this skill is invoked, first check if a profile exists at `~/.claude-job-profile.json`.

**If NO profile exists**, say:

> Welcome to the Job Application Assistant! I'll help you fill out job applications on LinkedIn, Greenhouse, and Workday.
>
> First, I need to set up your profile. This is a one-time process — your information will be saved for future applications.
>
> **Please provide the path to your resume file** (PDF, DOCX, or TXT).
>
> For example: `~/Documents/resume.pdf` or `/Users/you/Desktop/MyResume.pdf`

Then wait for the user to provide the path before proceeding with profile extraction.

**If a profile DOES exist**, say:

> Welcome back! Your profile is loaded from `~/.claude-job-profile.json`.
>
> **Provide a job URL** and I'll help you apply. For example:
> - LinkedIn: `https://www.linkedin.com/jobs/view/123456789`
> - Greenhouse: `https://boards.greenhouse.io/company/jobs/123`
> - Workday: `https://company.wd5.myworkdayjobs.com/jobs/job/123`
>
> Or say **"reset profile"** if you want to update your information from a new resume.

---

## Required Input

- **Resume file path**: Path to your resume (PDF, DOCX, or TXT format)
- **Job URL**: LinkedIn job posting or direct application link

## Profile Storage

Your extracted profile is stored at `~/.claude-job-profile.json` for reuse across sessions. Run with `--reset-profile` to re-extract from resume.

---

## Dual-Tool Architecture

This skill uses **two browser automation tools** together because each has capabilities the other lacks:

| Capability | Chrome MCP | Playwright MCP |
|---|---|---|
| Authenticated sessions (LinkedIn) | Yes | No — runs a separate browser with no login |
| File uploads | No — opens OS file picker that can't be controlled | Yes — `setInputFiles` and `browser_file_upload` |
| Iframe interaction | Limited — can't reliably reach iframe content | Yes — snapshots include iframe elements transparently |
| JavaScript injection | Yes — `javascript_tool` | Yes — `browser_evaluate` and `browser_run_code` |
| Dropdown/combobox interaction | Limited | Yes — `browser_fill_form` and `browser_click` |

### Tool Routing Rules

**Use Chrome MCP (`mcp__claude-in-chrome__*`) for:**
- LinkedIn navigation (requires logged-in session)
- Any site requiring authentication
- JavaScript injection to capture external URLs
- Reading LinkedIn job details

**Use Playwright MCP (`mcp__plugin_playwright_playwright__*`) for:**
- External application portals (Greenhouse, Ashby, Lever, Rippling, Workday)
- File uploads (resume, cover letter)
- Forms embedded in iframes
- Dropdown and combobox interaction
- Any form filling on non-authenticated pages

### The Handoff Pattern

1. **Chrome MCP** navigates to the LinkedIn job listing (authenticated)
2. **Chrome MCP** clicks Apply — if it's an external application, captures the external URL via JavaScript interception
3. **Playwright MCP** opens the captured URL and fills the form, uploads files, and submits

---

## LinkedIn → External URL Extraction

When a LinkedIn job uses an external application portal, the Apply button opens a new tab. Use JavaScript interception to capture that URL.

### Pattern A: Button with `window.open` (most common)

```javascript
// Step 1: Set up interceptor in Chrome MCP
var originalOpen = window.open;
window.open = function(url, target, features) {
  document.title = 'CAPTURED:' + url;
  var w = originalOpen.call(window, url, target, features);
  return w;
};

// Step 2: Click the Apply button
document.querySelector('button.jobs-apply-button').click();

// Step 3: Read document.title — it will start with "CAPTURED:" followed by the URL
```

### Pattern B: Link with `target="_blank"` (fallback)

```javascript
// Remove target="_blank" so it navigates in the same tab
var link = document.querySelector('a.jobs-apply-button');
link.removeAttribute('target');
link.click();
// Read the URL from the Chrome tab after navigation
```

After capturing the URL, pass it to Playwright MCP via `browser_navigate`.

---

## Workflow

### Phase 1: Profile Setup

If no profile exists at `~/.claude-job-profile.json`, or if the user requests a reset:

1. **Read the resume file** using the Read tool
2. **Extract structured data** into these categories:
   - `firstName`, `lastName`
   - `email`, `phone`
   - `location` (city, state, country, zip)
   - `linkedInUrl`, `portfolioUrl`, `githubUrl` (if present)
   - `workHistory[]`: array of { company, title, startDate, endDate, current, description }
   - `education[]`: array of { school, degree, field, startDate, endDate, gpa }
   - `skills[]`: array of skill strings
   - `resumePath`: absolute path to the resume file on disk
3. **Present extracted data to user** for review and correction
4. **Save confirmed profile** to `~/.claude-job-profile.json`

### Phase 2: Application Filling

1. **Load profile** from `~/.claude-job-profile.json`
2. **Determine starting point**:
   - If URL is LinkedIn (`linkedin.com`) → use **Chrome MCP** to navigate (authenticated)
   - If URL is a direct external portal → skip to step 5 with **Playwright MCP**
3. **Chrome MCP**: Navigate to LinkedIn job listing, click Apply
4. **Chrome MCP**: If external portal, capture URL using JS interception pattern (see above)
5. **Playwright MCP**: Navigate to external URL with `browser_navigate`
6. **Playwright MCP**: Detect platform, read form structure with `browser_snapshot`
7. **Playwright MCP**: Fill fields with `browser_fill_form`, handle dropdowns with `browser_click`
8. **Playwright MCP**: Upload resume with `browser_file_upload` or `browser_run_code` using `setInputFiles`
9. **Playwright MCP**: Take snapshot to verify all fields are filled correctly
10. **Present summary** to user showing all filled values, wait for explicit confirmation
11. **Playwright MCP**: Click Submit

---

## Platform-Specific Guidance

### LinkedIn Easy Apply

**Characteristics:**
- Modal-based multi-step wizard
- Usually 2-5 steps: Contact Info → Resume → Additional Questions → Review
- Has progress indicator at top

**Approach:**
1. Click "Easy Apply" button to open modal
2. Use `read_page` on each step to identify fields
3. Common fields:
   - Phone number (often pre-filled from LinkedIn)
   - Resume upload (use `upload_image` tool with resume path)
   - Work authorization questions (dropdowns)
   - Custom screening questions (varies by employer)
4. Click "Next" to advance, "Review" on final step
5. Screenshot the review page before "Submit application"

**Field patterns to look for:**
- `input[name*="phone"]` - Phone number
- `input[type="file"]` - Resume upload
- `select`, `[role="listbox"]` - Dropdown questions
- `[role="radio"]`, `[role="checkbox"]` - Multiple choice

### Greenhouse

**Characteristics:**
- Single long-form page with sections
- Clear field labels
- Often has "Add another" for work history/education
- **Frequently embedded in iframes** on company career sites — Playwright handles this transparently

**Approach (use Playwright MCP):**
1. Navigate to URL with `browser_navigate`
2. Use `browser_snapshot` to read full form — iframe content appears with `f54eXX` refs
3. Fill from top to bottom using `browser_fill_form`
4. **Phone country code**: Click the country code toggle → select "United States: +1" from the listbox → the phone field auto-formats with +1 prefix
5. For work history sections:
   - Fill most recent position
   - Click "Add another" if form allows and user has more history
6. Education section similar pattern
7. Handle custom questions at bottom
8. Upload resume via `browser_file_upload` or `browser_run_code` with `setInputFiles`
9. Single "Submit Application" button at end

**Field patterns:**
- Standard `<input>` and `<select>` elements
- `#first_name`, `#last_name`, `#email`, `#phone` common IDs
- `.field