By name: job-ranker description: | Scrape jobs from a company careers page and rank them by fit for the user. Use when user provides a company careers URL and wants jobs analyzed and ranked. Outputs a CSV with jobs ranked from TOP PICK (best fit) to STRETCH (possible but harder). Filters to US-based roles only. Reads user's personal_details.md for background matching.
Job Ranker Skill
Scrape jobs from a company careers page, filter to US roles, and rank by fit.
Workflow
- Navigate to the provided careers URL
- Extract all job listings (title, location, URL, requirements if visible)
- Filter to US-based roles only (Remote US, or US cities/states)
- Read user's background from
references/personal_details.md - Rank each job using the matching criteria below
- Output CSV file:
ranked_jobs/{company}_jobs_ranked.csv
CSV Output Format
Rank,Job Title,Location,Experience Required,Fit Category,Match Score,Key Matches,Gaps,Job URL
1,ML Engineer,Remote US,3-5 years,TOP PICK,85%,"Python, PyTorch, LLMs",None,https://...
Fit Categories
| Category | Match Score | Meaning |
|---|---|---|
| TOP PICK | 80-100% | Strong match, high confidence |
| HIGH | 60-79% | Good match, worth applying |
| STRETCH | 40-59% | Some gaps but possible |
| SKIP | <40% | Poor fit, don't apply |
Matching Criteria
User's Core Competencies
Read from references/personal_details.md:
- Primary programming languages
- Cloud platforms (AWS, GCP, Azure)
- Frameworks and tools
- Years of experience
- Work authorization/visa status
Auto-SKIP Rules (Do NOT include in CSV)
- Requires security clearance (if you can't get it)
- Requires more years than you have
- Staff/Principal/Director level (if too senior)
- Specialized tools you don't know
Scoring Heuristics
- +20% if role matches primary skills
- +15% if experience requirement matches your level
- +10% if remote or preferred location
- -15% if requires slightly more experience than you have
- -25% if requires significantly more experience
- -10% for each major required skill not in your background
Execution Notes
- Use browser automation (chrome tools) to navigate and scrape
- If careers page has filters, apply: Location=United States, remote
- If pagination exists, scrape all pages
- Extract job details from listing or click into each job if needed
- Save CSV to
ranked_jobs/directory - Report summary: total jobs found, jobs after filtering, breakdown by category
Usage Examples
User: "Rank jobs at https://amazon.jobs/en/search"
Action: Navigate, scrape all US software/ML roles, rank by fit, save to ranked_jobs/amazon_jobs_ranked.csv
User: "Find the best jobs for me at Apple" Action: Navigate to Apple careers, apply filters, rank all matches, save CSV
Output Summary Format
After completing the ranking, report:
Job Ranking Complete for [Company]
Total jobs found: X
After US filter: Y
After skip filter: Z
Breakdown:
- TOP PICK (80-100%): N jobs
- HIGH (60-79%): N jobs
- STRETCH (40-59%): N jobs
- SKIP (<40%): N jobs (excluded)
Top 5 Recommendations:
1. [Job Title] - [Match Score] - [Location]
2. ...
Saved to: ranked_jobs/{company}_jobs_ranked.csv