CVOps Generator

English | 中文

CVOps is an engineering wrapper around renderCV and CUE, designed to provide a structured and programmable way to generate high-quality resumes.

Why this Wrapper?

While renderCV is powerful, maintaining complex or multiple resumes in raw YAML can be repetitive and error-prone. This project offers a battle-tested engineering template that serves as a proven starting point for your resume:

1. Logic > Config (CUE vs YAML): Use variables, loops, and schemas to enforce consistency. Define your education or contact info once and reuse it across multiple resume versions (e.g., English/Chinese, specialized roles) without copy-pasting.
2. Engineering > Script: Instead of manually managing Python versions and LaTeX dependencies, this repo provides a fully configured Dev Container. It just works.
3. Systematic Design: Strict separation of Content (src/content), Design (src/design), and Locale (src/locale) means you can change your layout or language without rewriting your life story.

Setup

Important

Dev Container Support Only

We strongly suggest (and essentially only support) using this project within a Dev Container. The provided Dev Container comes pre-configured with all necessary tools (uv, cue, just, etc.) to get you started immediately without environment hell.

1. Open this repository in VS Code.
2. When prompted, reopen the folder in the Dev Container.
3. Wait for the container to build and initialization scripts to run.

Usage

Building a Resume

Use the just command runner to build your resume.

just build {template name}

Note: The {template name} is NOT the filename. It corresponds to the root-level object name defined in your CUE files within src/content/.

For example, if you have a definition like example_cn: ... in src/content/example_cn.cue, you would run:

just build example_cn

Creating New Resumes

• Location: All resume content must be written in src/content/.
• Best Practice: It is highly recommended to separate different resume contents into different .cue files (e.g., resume_en.cue, resume_cn.cue) for better organization.
• Root Object: Ensure each specific resume has a unique top-level key (which becomes your template name).

Example

1. Create src/content/my_resume.cue:

   package content
   
   my_awesome_resume: {
       // ... your resume content ...
   }

2. Build it:

   just build my_awesome_resume

---

Agent Skills (Multi-Agent Skills)

This project uses a shared skills/ directory to bundle Prompt + resources for coding agents. It is symlinked into common agent locations (.codex/skills, .claude/skills, .cursor/skills, .gemini/skills, .agent/skills, .github/skills) so different tooling can discover the same skills without duplication.

Included skills:

• resume-generate: read a JD and master data to produce cue-src/content/generated/{target}.cue.
• resume-audit: verify generated resume claims against master data.
• (Removed) resume-sync: master data sync is intentionally handled without an LLM for now.

Example usage (Codex CLI/IDE): $resume-generate, $resume-audit.

---

Vision & Roadmap: The "Agentic" Resume

Goal: High-quality mass application with minimized effort.

We are evolving this project from a static generator into an Agentic Career Assistant. The core philosophy is to decouple your "Experience Data" from the "Resume View".

The Concept

1. Atomic Assets: You don't write a resume; you build a database of Atomic Experiences. Each experience is a structured data block (Context, Problem, Method, Result) stored in your Experience Bank.
2. Dynamic Optimization: When you apply for a job, you feed the Job Description (JD) and your Experience Bank into the Agent.
3. Tailored Output: The Agent selects the most relevant atomic experiences, rewrites them to match the keywords and tone of the JD, and assembles a perfectly tailored resume in an instant.

Why?

• For Generalists: Highlight your technical depth for engineering roles and your leadership scope for management roles—using the same source of truth.
• For Freelancers: Generate specific proposals/profiles for every different client requirement.
• For Everyone: Break the trade-off between quantity (mass applying) and quality (tailored content).