Resume Autopilot — Claude Code Resume Tailoring Agent

An automated, ATS-optimized resume tailoring system built on Claude Code. Paste a job description, get a tailored, quality-gated, one-page PDF — without hand-editing LaTeX.

This repo ships with a fictional sample persona (“Alex Morgan”, a Data Engineer) so the whole pipeline runs end-to-end out of the box and compiles to a real demo PDF. Swap in your own details to make it yours.

New in v3.0.0: cover letter generation. A sixth skill (resume-cover-letter) writes a tone-gated, truth-checked one-page cover letter as a designed PDF — opt-in per role, reusing the resume’s analysis. See CHANGELOG.md.

Why this exists: sharing is caring. Job hunting is exhausting enough without spending an hour tweaking your resume for every posting. If tech can make that easier, everyone should have it — so here it is. Use it, fork it, pass it on.

---

The problem

Tailoring a resume to each job posting is the single highest-leverage thing a job seeker can do for ATS (applicant tracking system) match rates — and the most tedious. For every application you have to:

1. Read the JD and figure out which keywords it actually weights
2. Decide what to add, where, and what to cut to stay one page
3. Edit LaTeX without breaking the layout
4. Sanity-check metrics, buzzwords, length, and formatting
5. Recompile and track what you sent where

Doing this by hand for 20+ applications a week is unsustainable. Doing it with a naive “rewrite my resume” LLM prompt burns tokens and produces inconsistent, often-fabricated output.

This repo is the engineered answer: a deterministic-script-first, LLM-second pipeline driven by a Claude Code agent, with hard quality gates and a fixed LaTeX template that the agent is forbidden to break.

---

Demo

A fully-worked example ships in jobs/example_company/ — the persona’s resume tailored to a fictional Cloudpeak — Data Engineer posting:

The keyword script scored the master resume at 81% against this JD; after the agent executed the planned changes, the tailored resume scored 94% — passing all quality gates and staying on one page.

$ python scripts/keyword_match.py jobs/example_company/jd.txt master-1page/resume.tex

Match score: 81%  (26/32 known JD keywords)