Job Search Assistant

A Claude Code skill for searching LinkedIn jobs with a focus on finding positions where you have network connections or hiring managers are listed.

---

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 Search Assistant! I'll help you find jobs on LinkedIn where you have connections or hiring managers are listed.

First, I need to set up your profile so I can suggest relevant search terms based on your experience.

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, analyze it and present suggestions:

Welcome back! I found your profile at ~/.claude-job-profile.json.

Based on your resume, here's what I suggest for your job search:

Keywords: [Most recent job title], [Related titles based on skills] Location: [City, State from profile] Experience Level: [Calculated level] ([X] years of experience)

Would you like to:

1. Search with these suggestions
2. Modify the keywords, location, or other filters
3. Reset your profile from a new resume

Just let me know how you'd like to proceed!

---

Required Input

• Resume file path: Only if no profile exists at ~/.claude-job-profile.json

Auto-Inferred from Profile

These values are automatically extracted from the user's profile and presented for confirmation:

• Keywords: Suggested from job titles in workHistory and skills array
• Location: From profile.location (city, state)
• Experience Level: Calculated from total years in workHistory
  • 0-2 years: Entry level
  • 2-5 years: Associate
  • 5-10 years: Mid-Senior level
  • 10+ years: Director/Executive

Optional Filters

User can specify or modify:

• Work Type: remote, hybrid, onsite, or all
• Posted Within: 24 hours, past week, past month, or any time
• Max Results: Number of jobs to analyze (default: 10, max: 25)

---

Profile Storage

The profile is stored at ~/.claude-job-profile.json and shared with the /job-apply skill. Run /job-search reset profile to re-extract from a new resume.

Results Storage

Search results are automatically saved to ~/.claude-job-searches/search-{timestamp}.json. This enables:

• Review of past searches
• Easy reference for follow-up actions
• Structured data for further analysis

---

Workflow

Phase 1: Profile & Keyword Setup

1. Check for existing profile at ~/.claude-job-profile.json
2. If no profile exists:
   • Request resume file path from user
   • Read the resume file using the Read tool
   • Extract structured data into profile format (same as job-apply skill):
     • firstName, lastName, email, phone
     • location (city, state, country, zip)
     • linkedInUrl, portfolioUrl, githubUrl
     • workHistory[]: array of { company, title, startDate, endDate, current, description }
     • education[]: array of { school, degree, field, startDate, endDate, gpa }
     • skills[]: array of skill strings
   • Present extracted data to user for review
   • Save confirmed profile to ~/.claude-job-profile.json
3. Analyze profile to generate search suggestions:
   • Extract job titles from workHistory (most recent first)
   • Identify key skills from skills array
   • Generate related/adjacent job titles
   • Calculate experience level from work history dates
   • Extract location from profile
4. Present suggestions to user with all inferred values
5. Let user confirm or modify keywords, location, experience level
6. Collect additional preferences: work type, recency filter
7. Confirm final search parameters before proceeding

Phase 2: LinkedIn Navigation

1. Ensure browser context exists using tabs_context_mcp
2. Create new tab if needed using tabs_create_mcp
3. Navigate to LinkedIn Jobs: https://www.linkedin.com/jobs/
4. Verify user is logged in:
   • Use read_page to check for profile menu or sign-in buttons
   • If not logged in, STOP and instruct user: "Please log into LinkedIn in this browser tab, then let me know when you're ready."
5. Build search URL with parameters (see LinkedIn URL Parameters below)
6. Navigate to search URL
7. Wait for results to load using computer with wait action (2-3 seconds)

Phase 3: Job Extraction

For each job listing (up to maxResults):

1. Read job cards from results list using read_page
2. Click into job detail page using computer with left_click on job title
3. Wait for detail page to load (1-2 seconds)
4. Extract job information:
   • Title, company name, location
   • Posted date, applicant count
   • Work type (Remote/Hybrid/On-site)
   • Apply method (Easy Apply vs external)
5. Look for connection indicators:
   • Use find with query: "connections work here" or "connections at"
   • Extract connection count and names if visible
6. Look for hiring team section:
   • Use find with query: "hiring team" or "Meet the hiring team"
   • If found, extract hiring manager name, title, and profile URL
7. Store extracted data in results array
8. Navigate back to results list or click next job card
9. Add delay between jobs (2-3 seconds to avoid rate limiting)

Phase 4: Results Processing

1. Sort results by priority:
   • Priority 1 (Highest): Jobs with hiring manager listed
   • Priority 2: Jobs with 1st-degree connections
   • Priority 3: Jobs with 2nd-degree or alumni connections
   • Priority 4: Jobs with no network advantage
2. Create results directory if needed: ~/.claude-job-searches/
3. Save results to JSON file: ~/.claude-job-searches/search-{ISO-timestamp}.json
4. Display formatted summary in terminal (see Output Format below)
5. Inform user of saved file location

---

LinkedIn URL Parameters

Base URL: https://www.linkedin.com/jobs/search/

Parameter	Purpose	Values
keywords	Search terms	URL-encoded string
location	Geographic area	City, state, or country name
f_WT	Work type	1 (on-site), 2 (hybrid), 3 (remote)
f_TPR	Time posted	r86400 (24h), r604800 (week), r2592000 (month)
f_E	Experience level	1 (intern), 2 (entry), 3 (associate), 4 (mid-senior), 5 (director), 6 (executive)
f_JIYN	In your network	true (shows jobs at companies with connections)
f_AL	Actively hiring	true
f_EA	Easy Apply only	true
sortBy	Sort order	DD (date), R (relevance)

Example URL:

https://www.linkedin.com/jobs/search/?keywords=Senior%20Software%20Engineer&location=San%20Francisco%2C%20CA&f_WT=2%2C3&f_E=4&f_TPR=r604800&sortBy=DD

---

Chrome MCP Tool Usage

Checking Authentication

Use read_page to look for:
- Profile menu in top navigation (user is logged in)
- "Sign in" or "Join now" buttons (user is NOT logged in)

If not logged in, stop and inform user.

Reading Job Listings

Use read_page with filter: "interactive" to see job cards
Each job card typically contains:
- Job title (clickable link)
- Company name
- Location
- Posted date
- Easy Apply badge (if applicable)

Finding Connections

After clicking into job detail:
1. Use find with query: "connections work here"
2. Or use find with query: "connections at this company"
3. Look for section showing connection profile pictures/names

Finding Hiring Manager

After clicking into job detail:
1. Use find with query: "Meet the hiring team"
2. Or use find with query: "hiring manager"
3. If found, extract:
   - Name (usually in heading or profile card)
   - Title (usually below name)
   - Profile link (href on the card)

Scrolling for More Results

LinkedIn uses infinite scroll:
1. Use computer with action: "scroll", scroll_direction: "down"
2. Wait 2-3 seconds for new results to load
3. Use read_page to see newly loaded job cards
4. Repeat until desired count or no more results

---

Output Format

Terminal Display

=== LinkedIn Job Search Results ===
Search: "Senior Software Engineer" in San Francisco, CA
Filters: Remote/Hybrid | Past Week | Mid-Senior Level
Found: 12 jobs analyzed

Saved to: ~/.claude-job-searches/search-2026-01-06T10-30-00.json

-----------------------------------------------------------
1. [HIRING MANAGER] Senior Software Engineer
   Company: Acme Corp
   Location: San Francisco, CA (Remote)
   Posted: 2 days ago | 45 applicants

   Hiring Manager: Jane Smith (Engineering Manager)
   Connections: 2 (John Doe, Sarah Lee)

   Apply: Easy Apply
   URL: https://linkedin.com/jobs/view/123456
-----------------------------------------------------------

2. [CONNECTIONS] Staff Software Engineer
   Company: TechStart Inc
   Location: Palo Alto, CA (Hybrid)
   Posted: 1 day ago | 23 applicants

   Connections: 5 (including 2 Stanford alumni)

   Apply: Easy Apply
   URL: https://linkedin.com/jobs/view/789012
-----------------------------------------------------------

... (additional results)

=== Summary ===
- 2 jobs with hiring managers listed
- 4 jobs with 1st-degree connections
- 3 jobs with alumni connections
- 3 jobs with no network advantage

JSON File Schema

{
  "searchMetadata": {
    "keywords": "Senior Software Engineer",
    "location": "San Francisco, CA",
    "filters": {
      "workType": ["remote", "hybrid"],
      "postedWithin": "week",
      "experienceLevel": "mid-senior"
    },
    "timestamp": "2026-01-06T10:30:00Z",
    "totalJobsAnalyzed": 12
  },
  "jobs": [
    {
      "id": "123456",
      "url": "https://linkedin.com/jobs/view/123456",
      "title": "Senior Software Engineer",
      "company": {
        "name": "Acme Corp",
        "linkedinUrl": "https://linkedin.com/company/acme-corp"
      },
      "location": "San Francisco, CA",
      "workType": "remote",
      "postedDate": "2 days ago",
      "applicantCount": 45,
      "applyMethod": "easyApply",
      "hiringManager": {
        "name": "Jane Smith",
        "title": "Engineering Manager",
        "profileUrl": "https://linkedin.com/in/janesmith"
      },
      "connections": [
        {
          "name": "John Doe",
          "title": "Software Engineer",
          "connectionDegree": 1,
          "profileUrl": "https://linkedin.com/in/johndoe"
        }
      ],
      "connectionSummary": {
        "total": 2,
        "firstDegree": 2,
        "secondDegree": 0,
        "alumni": 0
      },
      "priority": "high"
    }
  ],
  "summary": {
    "withHiringManager": 2,
    "withFirstDegreeConnections": 4,
    "withAlumniConnections": 3,
    "noNetworkAdvantage": 3
  }
}

---

Safety Rules

1. Never enter credentials - If login is required, stop and instruct user to log in manually
2. Never click Apply - This skill is for searching only, not applying
3. Never create accounts - Stop and inform user if account creation is required
4. Respect rate limits - Add 2-3 second delays between page loads
5. Limit result count - Maximum 25 jobs per search to avoid excessive scraping
6. Handle errors gracefully - If a job page fails to load, skip and continue

---

Example Invocations

New user (no profile):

User: /job-search
Claude: Welcome! Please provide your resume path to get started.
User: ~/Documents/resume.pdf
Claude: [Parses resume, presents suggestions]
        Based on your resume, I suggest searching for "Product Manager" in Seattle, WA...

Returning user:

User: /job-search
Claude: Welcome back! Based on your profile, I suggest:
        Keywords: "Senior Software Engineer", "Staff Engineer"
        Location: San Francisco, CA
        Experience: Mid-Senior (7 years)

        Ready to search, or would you like to adjust?
User: Search for remote only
Claude: [Executes search with remote filter, displays results]

With specific request:

User: /job-search for machine learning roles in NYC
Claude: I'll search for machine learning roles in NYC. Based on your 5 years of experience,
        I'll filter for Mid-Senior level. Any preference for remote/hybrid/onsite?