AI Job Search is a job application workflow built on Claude Code. You copy the repository, build a profile of your experience and writing style, and then use slash commands to search job boards, judge how well each posting fits, draft a tailored CV and cover letter, and prepare for interviews. All data stays on your own machine.
The core loop has three steps. Setup builds the profile from a folder of career documents (CV, LinkedIn export, diplomas, references, past applications), a pasted CV, or a guided interview. Scrape searches job portals, removes duplicates, and lists matches by fit, and Rank batch-scores a large haul in parallel against five evaluation dimensions. Apply evaluates fit, drafts the CV and cover letter in LaTeX, has a second reviewer agent research the company and critique the drafts, revises, and compiles the PDFs.
The compile step is a verification loop: the agent reads the rendered pages and iterates until the CV is exactly two pages and the cover letter exactly one, then checks the CV's text layer the way an applicant tracking system parses it and reports keyword coverage. Claims are checked against the profile, and gaps are left visible rather than filled with invented experience. Job postings are treated as untrusted input.
Further commands prepare stage-specific interview packs with mock interviews, record outcomes and draft follow-ups, read Gmail for status changes and propose tracker updates for approval, publish a read-only pipeline view to Notion, analyze skill gaps into a learning plan, and generate an offline HTML dashboard. The bundled portal search tools target Danish job boards plus LinkedIn and a multi-market tech aggregator, and a command generates search skills for other markets' boards. Custom CV and cover letter templates in LaTeX or Typst can replace the stock ones.
Features
- Profile building: imports career documents, a pasted CV, or a guided interview into a structured candidate profile
- Job portal search: searches several boards, deduplicates results, and sorts them by fit
- Batch ranking: parallel agents score scraped postings and return a shortlist with strengths, gaps, and deadline flags
- Drafter-reviewer pipeline: tailored CV and cover letter drafts critiqued by a second agent that researches the company
- PDF verification loop: iterates on LaTeX until page counts, layout, and fonts meet the template rules
- ATS check: inspects the CV text layer as a parser sees it and reports keyword coverage without stuffing
- Interview prep: stage-specific prep packs, company research, STAR example mapping, and mock interviews
- Outcome tracking: archives submitted materials, records results, and drafts follow-ups without sending them
- Gmail and Notion sync: proposes status updates from email for approval and publishes a read-only Notion view
- Skill gap analysis: a prioritized gap heatmap and learning plan from tracked and ranked postings
- Portal generator: scaffolds and test-runs a search skill for a job board in your own market
- Custom templates: registers your own LaTeX or Typst CV and cover letter templates
