Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/SudharakaA/ai-job-search/llms.txt

Use this file to discover all available pages before exploring further.

/scrape is your recurring job search command. It reads your configured search queries, runs them against multiple job portals in parallel, filters out anything you have already seen or applied to, scores each new posting against your profile, and presents a ranked results table. Run it daily to catch new postings without sorting through duplicates. Deduplication state persists between runs in job_scraper/seen_jobs.json.

​
Syntax

/scrape
/scrape <focus-area>
/scrape broad
InvocationBehavior
/scrapeRuns the top 3 priority query categories from search-queries.md
/scrape data sciencePrioritizes queries from the matching focus category
/scrape broadRuns all query categories, not just the top 3

​
What /scrape does

1

Load state

Reads job_scraper/seen_jobs.json (created automatically on first run) and job_search_tracker.csv to build a list of already-seen URLs and already-applied company/role pairs.
2

Search portals

Runs WebSearch queries from search-queries.md targeting your configured job portals and geographic area. Searches are scoped to postings from the last 14 days. Parallel queries speed up this phase.
3

Fetch and parse

For each promising result, fetches the full posting page and extracts: job title, company, location, posting date, URL, key requirements, and application deadline.
4

Quick fit assessment

Scores each new posting against your profile at three levels: High (role directly involves your core skills), Medium (role is adjacent to your experience), or Low (role requires significant skills you lack). This is a fast signal, not the full evaluation that /apply runs.
5

Deduplicate and store

Skips any result whose URL, or company-plus-title combination, already exists in seen_jobs.json or job_search_tracker.csv. All fetched jobs (including skipped ones) are added to seen_jobs.json so they are not fetched again.
6

Present results

Displays new matches in a table sorted by fit, high first. High-match entries include 2-3 bullet points summarizing why the role fits, key requirements to check, and any flags.

​
Results table format

## New Job Matches - YYYY-MM-DD

Found X new positions (Y high, Z medium, W low match).

| # | Fit    | Title         | Company   | Location   | Deadline   | URL         |
|---|--------|---------------|-----------|------------|------------|-------------|
| 1 | High   | Data Engineer | Acme Corp | Copenhagen | 2025-08-01 | [Link](url) |
| 2 | Medium | ML Engineer   | Beta ApS  | Remote     | 2025-07-28 | [Link](url) |
High-match entries are followed by a brief highlights block explaining the fit, required skills worth reviewing, and any red flags.

​
Moving from a scrape result to an application

After presenting results, /scrape asks:
“Want me to evaluate any of these in detail? Just give me the number(s).”
Giving a number triggers the full /apply workflow on that posting, starting with a fit evaluation and moving to CV and cover letter drafting if you approve. You can also copy a URL from the table and run /apply <url> directly at any time.
If a portal blocks URL fetching when /apply tries to retrieve the posting, paste the full job description as the argument instead: /apply <pasted job description text>.

​
Configuring search queries

/scrape reads its queries from .claude/skills/job-scraper/search-queries.md. The file organizes queries into priority categories:
  • Priority 1: your strongest or most desired role direction
  • Priority 2: your domain expertise
  • Priority 3: adjacent roles you could pivot into
  • Priority 4: broader roles for a wider net
/setup generates this file from your profile during onboarding. You can edit it manually at any time, or re-run: 
/setup --section search
This reopens the job search configuration interview and regenerates the queries file without touching any other part of your profile.
The built-in portal skills target the Danish job market (Jobindex, Jobbank, Jobdanmark, Jobnet) plus LinkedIn public job listings. If you are in a different market, use /add-portal to generate a skill for your local job board, then add it to your search queries.

/setup

Configure search queries and set your target roles, skills, and location

/apply

Run the full application workflow on a posting URL or pasted description

/add-portal

Add a job portal search skill for your local market

/upskill

Analyze skill gaps across all tracked jobs

Build docs developers (and LLMs) love

Get started for freeTalk to us