Paste the prompt below into Claude (Claude Code in VS Code, or the chat at claude.ai). It won't ask you to fill in a blank template. It interviews you: it asks how you want your resume to look and read (your house style), it walks you through dumping your whole career into one master "super resume," and it asks what roles you're aiming for. When it's done, you have a standing system. From then on you just paste in a job description and say "tailor my resume," and out comes a focused one-page draft, a styled HTML version, and a PDF ready to send.
One master resume in. One job description in. One tailored result out, every time, in your own format, without re-explaining anything.
The prompt, copy everything in the box
You are going to help me build a personal resume-tailoring system in this project: the same setup that turns one master resume into a job-specific, one-page resume in minutes. Read this whole prompt first, then start. STEP 0 — figure out where you're running. - If you can create files and folders on disk (for example, you are Claude Code), build the real folder structure described below. - If you cannot write files (for example, you are a web chat), produce each file as a clearly labeled code block using the same file paths, so I can copy and save them myself. Tell me which mode you're in before you start. STEP 1 — interview me. Do NOT build anything yet, and do NOT dump every question at once. Ask about ONE area at a time, wait for my answer, ask brief follow-ups if my answer is thin, then move to the next area. Be a sharp interviewer, not a form. The four areas, in order: 1. HOUSE STYLE AND FORMATTING. How I want every resume to look and read. Cover: one page or two; section order (summary, experience, projects, education, skills); bullet style and whether bullets must fit on a single line; voice (active, metrics-first); and writing rules I want enforced every time. Offer these "no-AI-speak" defaults and let me accept or change them: no em dashes, use commas instead; no rule-of-three / "triples" in prose; no mechanically parallel bullet openings; unique starting verb for every bullet with a verb audit each run; plain language with concrete numbers. Capture anything I never want to see and anything that must always appear. 2. MY BACKGROUND — the super resume. Walk me through building one exhaustive master document that holds EVERYTHING: every role, project, and accomplishment, written from several angles, deliberately too long for one page. Ask probing questions to pull out what I'd undersell: scope, scale, metrics, dollar impact, what broke and how I fixed it, what I shipped, who I influenced. Go role by role. Push for numbers. This master file is the source of truth that every tailored resume is cut down from, so over-collect now. 3. TARGET ROLES AND EMPHASIS. What kinds of jobs I'm aiming for, the titles and seniority, and which of my strengths to foreground so tailoring has direction. Ask what I want a hiring manager to believe about me in ten seconds. 4. LINKS, CONTACT, AND MUST-INCLUDE FACTS. Name, contact line, LinkedIn, portfolio, GitHub, and anything that must always appear verbatim: fixed job titles, exact employment dates, certifications, work authorization. Confirm every link is a real working URL, never a placeholder. STEP 2 — once you have my answers, build the system: - `.claude/CLAUDE.md` — my house-style config. Write my formatting and writing rules here as a numbered, unambiguous checklist (the way I answered area 1), plus a short workflow section telling you that whenever I add a job description and ask you to tailor, you must: read the JD, read the super resume, apply ALL the rules, write `outputs/Resume_Tailored.md`, then render `outputs/Resume_Tailored.html` and a PDF. This file is the whole trick: because you read it automatically, every future draft comes out in my style without me re-explaining it. - `inputs/master/Resume_Super.md` — my exhaustive super resume, built from area 2. - `inputs/jd/` — empty, this is where I'll drop each target job description. Add a one-line README in it explaining that. - `outputs/` — empty, where tailored results land. - A styled HTML template for the rendered resume that matches the house style from area 1 (clean, one page, print-friendly with tight margins). Wire the workflow so the tailored Markdown fills this template to produce `outputs/Resume_Tailored.html`. - PDF: explain both paths and set up whichever fits my environment. Simple path for anyone: open the HTML and use the browser's Print, then Save as PDF. Automated path for Claude Code: a small script that prints the HTML to `outputs/Resume_Tailored.pdf` headlessly (Playwright or `wkhtmltopdf`); offer to add it. STEP 3 — finish by showing me how to use it from now on, in three lines: (1) drop a job description file into `inputs/jd/`, (2) tell you "tailor my resume to this JD," (3) get back a tailored `.md`, a styled `.html`, and a `.pdf`. Then offer to do a first live run: ask me to paste a real job description and produce my first tailored resume end to end so I can see the whole flow before I rely on it. Begin now with STEP 0, then STEP 1, area 1. Ask me your first question.
What happens after you paste it
- Claude tells you whether it's building real files or handing you code blocks, then starts the interview, one area at a time.
- You answer. The super-resume part is the longest, and it's worth it: that one master document is what every future tailoring pulls from.
- Claude builds your
.claude/CLAUDE.mdconfig, your super resume, and the project folders, then offers a first live run. - From then on the loop is just: paste a job description into
inputs/jd/, say "tailor my resume," and collect the Markdown, HTML, and PDF.
The point isn't the resume. It's that your format and rules live in a config file Claude reads on its own, so you never re-explain "make the bullets one line, no em dashes, one page" again. You set the rules once and reuse them for every job.