AI Job Search

Overview

AI Job Search is a structured, local workflow that turns Claude Code into a job application assistant. It covers self-profiling, fit evaluation, and a drafter-reviewer pipeline for tailored CVs and cover letters; the core workflow is language- and country-agnostic, while the shipped portal search skills target Denmark and can be extended for other markets.

Key Features

• Evaluates job postings against a candidate profile and returns fit scores, deal-breaker vetoes, and ranked shortlists.
• Drafts tailored CVs and cover letters in LaTeX through a drafter-reviewer workflow, then compiles and visually inspects the PDFs.
• Runs ATS parseability checks on the compiled CV text layer and verifies contact details, reading order, and keyword coverage without stuffing unsupported skills.
• Prepares candidates for interviews with company research, STAR question mapping, gap-aware bridge answers, and mock interviews.
• Tracks applications and outcomes, archives submitted materials, drafts follow-ups, and generates a self-contained offline HTML dashboard.
• Supports extension through job-portal search skills, custom CV or cover-letter templates, and configurable evaluation criteria.

Use Cases

• Job seekers who want a local, Claude Code-based workflow for evaluating postings and preparing tailored applications.
• Candidates targeting the Danish job market through the shipped Jobindex, Jobnet, Jobbank, and Jobdanmark search skills.
• Users adapting the framework to another job market with /add-portal or borrowing community portal skills.
• Career planners using skill-gap analysis and learning plans between applications.

Getting Started

• Fork or clone the repository, run python3 tools/install_portal_clis.py to install the portal CLI tools, start Claude Code, and use /setup to populate a profile.
• Use /scrape to search job portals, optionally /rank to batch-score postings, and /apply with a job URL or pasted job description to run the full application workflow.

Deployment & Requirements

• Claude Code CLI and either a Claude Pro/Max/Team subscription or Anthropic API credits.
• Python 3.10+.
• Bun for the job search CLI tools.
• A LaTeX distribution with lualatex and xelatex; minimal TeX installs may need extra packages.
• Optional pypdf for the ATS parseability check; Poppler pdftotext is a fallback.

Before You Adopt

• License: MIT. Review its terms before using, modifying, or distributing the project.
• The workflow depends on Claude Code, which has no free tier and requires a Claude subscription or Anthropic API credits.
• Job postings are treated as untrusted input, but the agentic defenses are instruction-level rather than a sandbox, so users should review fetched and written content on unfamiliar boards.
• The LinkedIn search skill is intended for personal use only, and automated access is against LinkedIn's Terms of Service, so volume should be kept low.
• The project is independent and is not affiliated with, endorsed by, sponsored by, or maintained by Anthropic.