Free 40-page Claude guide — download today
April 16, 2026Claude Skills Hubclaudecareer-opsai-agent

Career-Ops: The Open-Source AI Job Search Agent Built on Claude Code (34K Stars)

Career-Ops turns Claude Code into a full job search command center — auto-evaluates offers, generates tailored CVs, scans job portals, and tracks your pipeline. 34K GitHub stars. Here's how to set it up.

What Is Career-Ops?

Career-Ops is an open-source AI-powered job search system built on top of Claude Code. Created by Santiago Fernandez (@santifer), it turns your Claude Code CLI into a full job search command center.

The pitch: companies use AI to filter candidates. Career-Ops gives candidates AI to choose companies.

It's not a spray-and-pray tool. It's a filter that helps you find the few offers worth your time out of hundreds, then generates tailored materials for each one.

The numbers: 34,000+ GitHub stars. MIT licensed. Santiago used it personally to evaluate 740+ job listings, generate 100+ personalized CVs, and land a Head of Applied AI role.

What It Actually Does

Career-Ops is a single slash command (/career-ops) with multiple modes:

ModeWhat it does
Auto-PipelinePaste a job URL, get a full evaluation + tailored PDF CV + tracker entry
A-F Evaluation6-block analysis: role summary, CV match, level strategy, comp research, personalization, interview prep
ATS PDF GenerationKeyword-injected CVs customized per job description
Portal ScannerScans 45+ companies (Anthropic, OpenAI, ElevenLabs, Retool, n8n) across Greenhouse, Ashby, Lever, Wellfound
Batch ProcessingEvaluate 10+ offers in parallel using Claude sub-agents
Interview Story BankAccumulates STAR+Reflection stories across evaluations
Negotiation ScriptsSalary negotiation frameworks, geographic discount pushback, competing offer leverage
Dashboard TUITerminal UI to browse, filter, and sort your pipeline

The system is agentic: Claude Code navigates career pages with Playwright, evaluates fit by reasoning about your CV vs the job description (not keyword matching), and adapts your resume per listing.

How to Set It Up

Prerequisites

  • Claude Code installed and working
  • Node.js 18+
  • A markdown file with your CV (cv.md)

Quick Start

# Clone and install
git clone https://github.com/santifer/career-ops.git
cd career-ops && npm install
npx playwright install chromium   # For PDF generation

# Check setup
npm run doctor

# Configure your profile
cp config/profile.example.yml config/profile.yml
# Edit profile.yml with your details

# Add your CV
# Create cv.md in the project root with your CV in markdown

# Start Claude Code in the directory
claude

# Then customize:
# "Change the archetypes to backend engineering roles"
# "Translate the modes to English"
# "Add these 5 companies to portals.yml"

Key Insight: The System Learns You

Career-Ops isn't great on day one. It doesn't know you yet. Feed it context — your CV, your career story, your proof points, your preferences. The more you nurture it, the better the evaluations get. Think of it as onboarding a new recruiter: the first week they need to learn about you, then they become invaluable.

The 14 Skill Modes

Career-Ops ships with 14 modes as Claude Code skill files:

/career-ops                → Show all available commands
/career-ops {paste a JD}   → Full auto-pipeline
/career-ops scan           → Scan portals for new offers
/career-ops pdf            → Generate ATS-optimized CV
/career-ops batch          → Batch evaluate multiple offers
/career-ops tracker        → View application status
/career-ops apply          → Fill application forms with AI
/career-ops pipeline       → Process pending URLs
/career-ops contacto       → LinkedIn outreach message
/career-ops deep           → Deep company research
/career-ops training       → Evaluate a course/cert
/career-ops project        → Evaluate a portfolio project

The system is designed to be customized by Claude itself. Modes, archetypes, scoring weights, negotiation scripts — just ask Claude to change them.

How the Evaluation Works

You paste a job URL or description
        │
        ▼
┌──────────────────┐
│  Archetype       │  Classifies the role type
│  Detection       │
└────────┬─────────┘
         │
┌────────▼─────────┐
│  A-F Evaluation  │  Match analysis, gap identification,
│  (reads cv.md)   │  comp research, STAR stories
└────────┬─────────┘
         │
         ▼
  Score < 4.0? → Skip (don't waste your time)
  Score ≥ 4.0? → Generate tailored CV + add to tracker

The evaluation uses 10 weighted dimensions for scoring. It reasons about fit — not keyword matching. If the role scores below 4.0/5, the system recommends not applying. Your time is valuable.

Why This Matters for Claude Code Users

Career-Ops demonstrates what's possible with Claude Code skill files at scale. The 14 modes are essentially 14 specialized agents, each with its own skill file, sharing context through a pipeline tracker.

If you're building your own Claude Code workflows, Career-Ops is worth studying for the architecture patterns:

  1. Multi-mode skill files — one slash command routing to multiple specialized behaviors
  2. Persistent state — tracker file as single source of truth across sessions
  3. Batch sub-agents — parallel evaluation using claude -p workers
  4. Browser automation — Playwright integration for scanning live job portals
  5. PDF generation — creating polished documents from Claude Code output

These patterns apply to any domain where you need multiple specialized agents sharing state.

Combining Career-Ops with CLSkills Prompt Codes

Career-Ops handles the job search pipeline. Our prompt codes handle how you communicate during that pipeline:

  • Before a cover letter draft: prefix with /ghost to strip AI tells. Recruiters can spot ChatGPT cover letters instantly.
  • Before evaluating a job offer trade-off: prefix with L99 to force Claude to commit to a recommendation instead of listing pros and cons.
  • Before interview prep: prefix with /skeptic so Claude challenges your STAR stories before the interviewer does.
  • Before salary negotiation: prefix with /blindspots to surface the assumptions you're making about the offer.

The Combo Generator has paste-ready templates for common career tasks — try searching "cold email" or "negotiation" for stacks that compound.

Installation

Full setup guide: github.com/santifer/career-ops/blob/main/docs/SETUP.md

Quick version:

git clone https://github.com/santifer/career-ops.git
cd career-ops
npm install
npx playwright install chromium
npm run doctor
cp config/profile.example.yml config/profile.yml
# Edit profile.yml, add cv.md, then: claude

Requires Claude Code CLI. If you're new to Claude Code, start with our complete setup guide ($5) or the free skills library to get familiar with how skill files work.

Credit

Career-Ops is built and maintained by Santiago Fernandez (@santifer on GitHub, @santifer on X). MIT licensed. All credit for the system goes to him — we're writing about it because it's genuinely useful and demonstrates what Claude Code skill architecture looks like at scale.

If you use it and find value, star the repo and join the Discord community.

Related Resources

FAQ

Is Career-Ops free?

Yes. MIT licensed, fully open source. You need a Claude Code subscription (which requires an Anthropic API key or Claude Pro/Max plan) to run it.

Does it actually submit applications for me?

No. Career-Ops is human-in-the-loop by design. It evaluates, generates materials, and recommends — you decide and act. The system never submits an application without your explicit approval.

Can I use it for non-tech roles?

Yes. The archetypes are customizable — ask Claude to change them to match your field. Santiago built it for AI/ML roles but the evaluation framework is role-agnostic.

How does it compare to LinkedIn Easy Apply?

LinkedIn Easy Apply is spray-and-pray — apply to everything, hope something sticks. Career-Ops is the opposite: filter ruthlessly, apply to only the high-fit roles, and customize every application. The A-F scoring system means you spend time only on roles scoring 4.0+ out of 5.

Can I add my own companies to scan?

Yes. Edit portals.yml to add any company with a careers page. 45+ companies are pre-configured including Anthropic, OpenAI, ElevenLabs, Retool, and n8n.

What skill files does Career-Ops use?

Career-Ops uses 14 Claude Code skill files, one per mode. They live in the project's skill directory and are loaded by Claude Code at session start. If you want to understand how skill files work in general, see our guide to Claude Code skills or browse our 2,300+ skill library.

Want the full research library?

120 tested Claude prompt codes with before/after output and token deltas.

See the Cheat Sheet — $15