tailored-cv

Tailor your CV and cover letter to each job posting with an AI agent (Claude Code or OpenCode), and generate ATS-optimized PDFs ready to send.

The core idea is a professional profile that grows with use: whenever a posting asks for something that is not on record, the agent asks you, saves the answer and reuses it in future applications. It never invents data: everything in the CV comes from your profile or from something you confirmed.

How it works

Piece	What it does
.claude/skills/cv-tailor/	Reads the posting (link or text), matches it against your profile, asks you what is missing, writes the tailored CV and cover letter and generates the PDFs.
.claude/skills/profile-audit/	Reviews your profile against ATS best practices and asks you the questions needed to improve it.
docs/ats-guidelines.md	The ATS rules both skills apply.
cv/	Markdown → PDF build with Node and headless Chrome/Edge.
workspace.example/	Empty template for your personal workspace.
workspace/	Your workspace: profile, base CV, applications and generated PDFs. It is in .gitignore.

Getting started

Requirements: Node 18+ and Chrome or Edge installed.

git clone https://github.com/theviderlab/tailored-cv.git
cd tailored-cv
cp -r workspace.example workspace
cd cv && npm install

Then open the project with Claude Code and ask something like "build my profile from this CV" (pasting your CV) or straight away "tailor my CV to this posting: ". If workspace/ does not exist, the skill creates it from the template and guides you.

Usage

In Claude Code, pass the posting link to the skill:

/cv-tailor https://www.linkedin.com/jobs/view/1234567890

The agent then:

1. Reads the posting (if the link cannot be fetched, e.g. behind a login, it asks you to paste the text).
2. Compares it with your profile and classifies each requirement: already on record, a known gap, or unknown.
3. Asks you about anything unclear in a single message (e.g. "The posting asks for Kubernetes — do you use it?") and saves your answers to the profile.
4. Writes the tailored CV and cover letter, shows you what changed and where each fact comes from, and builds the PDFs once you approve.
5. Logs the application in workspace/applications.md.

The result lands in its own subfolder, one per posting:

workspace/output/<company>-<role>/
├── cv-<company>-<role>.md
├── cv-<company>-<role>.pdf
├── cover-letter-<company>-<role>.md
└── cover-letter-<company>-<role>.pdf

Other things you can ask: "update the Acme application to interview", "add to my profile that I use Terraform", or /profile-audit to polish the profile for ATS without a specific posting.

Your workspace

Profile (workspace/profile/)

The profile is the source of truth the CVs are built from. You don't have to fill it in upfront: it grows over time as you give the agent information, whether from your initial CV, the answers to its questions on each application, or anything you tell it directly. Every entry records where it came from and when.

File	What it stores
skills.md	Technical and management skills, with level and evidence
experience.md	Achievement bank per role (more bullets than fit in a CV)
projects.md	Projects, talks and community work
education.md	Degrees, certifications and languages
stories.md	Reusable anecdotes and arguments for cover letters
facts.md	Practical data: contact, location, availability, answers to application forms
gaps.md	What you confirmed you do not know or don't want to highlight, so it is not asked again
audit.md	State of the ATS audit (profile-audit): pending, resolved and dismissed items

workspace/cv-base.md and workspace/cover-letter-base.md are the default selection (up to two pages) that each application starts from; the profile can hold much more than fits in a CV.

Application log (workspace/applications.md)

A table with one row per application, most recent first: date, company, role, output folder, posting link, status and notes. cv-tailor adds the row when it builds the PDFs, and you update the status yourself or by asking the agent. Statuses:

Status	Meaning
?	PDFs generated, sending not confirmed yet
sent	Application sent
interview	In the interview process
offer	Offer received
rejected	Rejected
dropped	Decided not to send it

The log also lets the agent warn you when you already applied to a company, and suggest updating the base CV when the same addition keeps showing up across applications.

Building PDFs by hand

cd cv
npm run cv -- ../workspace/cv-base.md          # ATS-optimized CV
npm run cover -- ../workspace/cover-letter-base.md

Backing up your data (optional, recommended)

workspace/ stays out of this repo, so you can version it as an independent private repo:

cd workspace
git init
git add .
git commit -m "Initial profile"
gh repo create <your-user>/tailored-cv-workspace --private --source . --push

From then on, changes to the tool are committed at the root and changes to your data inside workspace/.

Safety net against leaks

The .githooks/pre-commit hook blocks any commit to the tool's repo that includes files from workspace/ or text matching the patterns in workspace/.private-patterns (your name, email, phone…; one per line). Enable it once per clone:

git config core.hooksPath .githooks

Language

The skills talk to you in your language; the profile is kept in English and the CV content follows the language of the posting. For a CV or cover letter in another language, start the markdown with front matter (--- / lang: es / ---): the PDF headings come from cv/i18n.json, where more languages can be added.