Claude Code Can Run Your Entire Job Search From the Command Line

A GitHub repo called AI Job Search turns Claude Code into a full job-hunting pipeline that runs on your own machine. It does not just scrape listings and hand you links. It searches job portals, scores how well each posting fits you, drafts a tailored CV and cover letter, runs a second agent to critique that draft, and preps you for the interview. The repo hit #1 Repository Of The Day on TrendShift, and if you have used Claude Code at all, you can get it running.

What the Repo Actually Does

The video frames show the README for a project titled "AI Job Search" with the tagline "The job search that runs on your machine." It is an AI-powered job application framework built on Claude Code. The pitch in the transcript is the differentiator: most job tools scrape postings and dump links on you. This one carries the process further. It evaluates your fit for each role, drafts a custom CV, and includes a reviewing agent that critiques what it produced before you send anything.

The key idea is that the whole thing runs locally. You fork the repo, fill in your profile, and Claude does the searching, ranking, writing, and prep work on your machine rather than through a hosted service.

The Workflow: Setup, Scrape, Apply

One frame lays out the pipeline as three parallel command tracks. It maps directly to how the framework is structured.

Command	What it does
/setup	Fill in your profile so your candidate files are ready
/scrape	Search job portals and present matches with fit ratings
/apply <url>	Evaluate fit, score and recommend, draft a tailored CV plus cover letter in LaTeX, then run a reviewer agent that critiques and revises before final output

The transcript calls out the reviewing agent specifically, and the frame confirms it: after the CV and cover letter are drafted, a reviewer agent critiques the work, revises, and produces the final output. That review loop is the part that separates this from a template filler.

Beyond these three, the repo ships more commands. The transcript mentions /interview, which preps you for the interview portion of a role. The repo also includes commands for result tracking, job ranking, gap analysis, and syncing your pipeline to tools like Notion and Gmail.

Built for Editing and Customization

The transcript stresses that the framework is easy to customize, and one frame backs this up with a "Customization" section titled "Which files to edit manually." Instead of running /setup, you can edit the profile files directly. The frame lists them:

File	What to change
CLAUDE.md	Your full profile: name, education, experience
01-candidate-profile.md	Structured version of your CV data
02-behavioral-profile.md	Your behavioral assessment or self-assessment
04-job-evaluation.md	Skill match areas, career goals, motivation
05-cv-templates.md	Profile statement templates for different roles
07-interview-prep.md	Your STAR examples from actual experience
search-queries.md	Job search queries for your skills and location

This is plain markdown you control. As the transcript puts it, any experience with Claude Code is enough to get it working.

The Regional Catch Worth Knowing

One frame notes the framework was built for the Danish market. The portal search tools cover Jobindex, Jobnet, Jobbank, and Jobdanmark, plus a country-agnostic LinkedIn tool. If you are not job hunting in Denmark, the portal layer is the piece you swap out. The frame states directly that the portal search skills "are designed to be swapped for your local job boards," and the repo provides an /add-portal command to generate scrapers for your own market. The career-guidance logic, fit scoring, CV framing, and salary benchmarking are portable. The scraping targets are not.

Key Takeaways

• AI Job Search is a Claude Code framework that runs the full application pipeline locally: search, fit scoring, CV and cover letter drafting, review, and interview prep.
• A separate reviewer agent critiques and revises the drafted CV and cover letter before you get the final output, which is the main thing that sets it apart from link-scraping tools.
• The workflow centers on three commands: /setup to build your profile, /scrape to find rated matches, and /apply <url> to draft and refine an application.
• Your profile lives in editable markdown files (CLAUDE.md, 01-candidate-profile.md, and others), so you can customize it directly without running the setup command.
• It ships wired to Danish job portals. Swap in your local boards with /add-portal if you are hunting elsewhere.

Resources

• MadsLorentzen/ai-job-search -- The GitHub repo shown in the video, an AI-powered job application framework built on Claude Code.
• Claude Code -- Anthropic's agentic command-line coding tool that the framework runs on.