LaTeX-obsessed agents that apply for jobs on your behalf

What it does

This is a Claude Code extension that encodes the entire job application workflow into slash commands. You populate a structured profile, then the system searches job portals, scores postings against your criteria, and for each target role drafts a CV and cover letter in LaTeX. A second “reviewer” agent critiques the drafts before final compilation.

The interesting bit

The project treats PDF layout as a first-class problem. The apply workflow compiles every CV and cover letter, inspects the rendered pages, and iterates on the LaTeX until the CV is exactly two pages with no orphaned headers and the cover letter fits exactly on one. It even relevance-weights content when cutting overflow, rather than blindly trimming oldest entries.

Key highlights

• Drafter-reviewer separation: one agent writes, a fresh agent researches the company and critiques, then the drafter revises.
• Relevance-weighted CV cutting scores each bullet by keyword match and uniqueness before deciding what to drop when space runs out.
• expand enriches your profile by scraping linked public sources (GitHub, Kaggle, syllabi) for competencies you forgot to list.
• upskill maps gaps between your profile and tracked postings, then builds a prioritized learning plan with time estimates.
• The core workflow is language- and country-agnostic, though the included job scrapers are hard-wired for the Danish market.

Caveats

• The job portal CLI tools are built specifically for Denmark (Jobindex, Jobnet, Jobbank, Jobdanmark), so you’ll need to build your own scrapers for other markets.
• It requires a local LaTeX distribution with lualatex and xelatex, plus Python and Bun, making the setup stack somewhat deep.
• The README notes that some job portals block automated access, forcing you to paste descriptions manually.

Verdict

Worth a look if you’re applying to dozens of roles and want machine assistance that still produces human-vetted, properly typeset output. Skip it if you don’t live in Denmark or aren’t willing to maintain a LaTeX toolchain.