convez.ai
SkillAutomaçãoIntermediário
RajkiranVS

Remote Job Hunter

RajkiranVS/openclaw-remote-job-hunter

Busca, pontua e entraga vagas remotas automaticamente com análise de skill gaps

Ver no GitHub
InicialPluginScore · de 0 a 100

Skill que automatiza a busca diária de vagas remotas em 10+ plataformas, pontuando candidatos com NLP contra seu currículo e identificando gaps de habilidades por frequência de mercado. Entrega resumo das melhores matches via WhatsApp através de agente OpenClaw.

Instalação

$npx convez-ai RajkiranVS/openclaw-remote-job-hunter

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
1
Forks
1
Acompanhando
0
Saúde
28%
Último commit
há 8 meses
Versão
—
Issues abertas
0
Contribuidores
1

Casos de uso

  • Monitoramento automático diário de vagas remotas sem intervenção manual
  • Scoring de compatibilidade job-candidate usando NLP e keyword matching
  • Identificação de skill gaps priorizados pela demanda de mercado
  • Notificação via WhatsApp de vagas de alto match (70%+)

Veredito

Solução robusta e bem arquitetada para automação end-to-end de job hunting. Combina scraping multi-plataforma, scoring NLP prático e integração com WhatsApp — ideal para quem busca eficiência em processo recorrente. Exige setup inicial (API keys, profile config) mas oferece ROI claro em tempo economizado.

README

A tradução para o português deste README ainda está na fila.

remote-job-hunter

Daily remote job search, NLP resume scoring, skill gap analysis, and WhatsApp delivery — fully automated.

version platforms domains license


What it does

Every day, automatically:

  1. 🔍 Scrapes 10+ job platforms — Remotive, RemoteOK, Jobicy, WeWorkRemotely, Himalayas, WorkingNomads, Adzuna (AU/NZ/SG/UK), Arbeitnow EU, and more
  2. 📊 Scores every job 0–100% against your resume using NLP keyword matching
  3. 🎯 Identifies skill gaps ranked by market frequency — tells you exactly what to learn
  4. 📱 Delivers a WhatsApp summary of top matches via OpenClaw agent
  5. 🌍 EU visa-sponsored fallback — automatically widens search when local supply is thin

Architecture

┌─────────────────────────────────────────────────────────┐
│                    remote-job-hunter                     │
│                                                         │
│  search.py          scorer.py           report.py       │
│  ──────────         ──────────          ──────────      │
│  10+ platforms  →   NLP scoring    →    MD report       │
│  EU fallback        Resume parse        Gap analysis    │
│  Adzuna API         Keyword match       WhatsApp msg    │
│  Dedup/filter       Title exclude                       │
└──────────────────────────┬──────────────────────────────┘
                           │
                    OpenClaw Agent
                    (Haiku-class LLM)
                           │
                    WhatsApp delivery
                    via cron schedule

Why it beats manual job searching

Manual searchremote-job-hunter
1–2 hours/day0 minutes/day
Miss jobs posted overnightCatches everything daily at 10 AM
No idea how well you matchExact % score per job
No skill gap visibilityTop gaps ranked by market frequency
Apply to everythingOnly surface strong matches (70%+)
Single platform10+ platforms including geo-targeted boards

Quick Start

1. Install dependencies

pip install requests python-docx pymupdf --break-system-packages

2. Set up API credentials

cp .env.example .env
# Edit .env and add your Adzuna credentials (free at developer.adzuna.com)
# ADZUNA_APP_ID=your_app_id
# ADZUNA_APP_KEY=your_app_key

3. Configure your profile

cp config/profile.template.json my-profile.json
# Edit my-profile.json with your name, resume path, target salary, domain

4. Run your first search

python3 src/main.py \
  --profile-config my-profile.json \
  --profile-meta profiles/ai-ml.json \
  --output daily_report.md

5. Read your results

cat daily_report.md

Supported Role Domains

DomainTarget Roles
ai-mlAI/ML Architect, MLOps Engineer, GenAI Engineer, Data Scientist, LLM Engineer
qa-automationTOSCA Lead, Test Architect, QA Automation Engineer, SDET, Test Lead
software-devFull Stack, Backend, Frontend, Software Engineer
devopsDevOps Engineer, Platform Engineer, SRE, Cloud Engineer

Sample Output

📊 Summary:
   Total jobs: 139
   🟢 80%+ match: 8

⚠️  Top skill gaps:
   - Playwright (13 jobs) — Medium term
   - TypeScript (8 jobs) — Medium term
   - Java (3 jobs) — Medium term

📋 Top matches:
   1. Senior QA Automation Engineer — Tech Aalto (100%) — AU
   2. Automation Tester — Hudson (100%) — AU
   3. Test Automation Engineer — Accenture (100%) — AU

Daily WhatsApp automation (OpenClaw)

Add this to your OpenClaw cron to get daily WhatsApp job summaries:

openclaw cron add --agent your-agent --name "Daily Job Search" \
  --cron "0 10 * * *" --tz "Asia/Kolkata" \
  --model "anthropic/claude-haiku-4-5-20251001" \
  --announce --channel whatsapp --to "+91XXXXXXXXXX" \
  --message "Run: cd /path/to/remote-job-hunter && python3 src/main.py \
    --profile-config my-profile.json \
    --profile-meta profiles/ai-ml.json \
    --output ~/daily_report.md. \
    Then read ~/daily_report.md and report the results."

How the scorer works

  1. Resume parsing — extracts text from .docx or .pdf, handles XML-split phrases
  2. Keyword expansion — pipe-separated skill lists ("API | UI | Mobile Testing") are expanded into full phrases
  3. Domain keyword matching — 80+ keywords per domain, curated per role family
  4. Title exclusion filter — hard-caps off-domain roles (e.g. DevOps jobs scoring high on cloud keywords for an AI/ML profile)
  5. Minimum signal threshold — jobs with fewer than 3 extractable skills are marked N/A rather than inflated

🌍 EU Job Search & Visa Sponsorship

Built specifically for professionals targeting Europe from outside the EU — India, APAC, LATAM, etc.

How it works

The pipeline runs in two stages:

Stage 1 — Global WFA search
  └── Remotive, RemoteOK, Jobicy, WeWorkRemotely, Himalayas, WorkingNomads
  └── Scores and ranks all matches

Stage 2 — EU Fallback (auto-triggered when Stage 1 < threshold)
  └── Arbeitnow  →  EU visa-sponsored roles across all of Europe
  └── Adzuna     →  AU, NZ, SG, UK markets (English-language, easy to apply)
  └── Merged, deduplicated, scored against your resume

Stage 2 fires automatically — you don't have to do anything. If your target market has thin supply that day, the pipeline widens the net without you noticing.

Enabling EU fallback

Add this to your profile JSON:

"eu_fallback": {
  "enabled": true,
  "min_wfa_threshold": 15,
  "allowed_regions": [
    "netherlands", "germany", "ireland", "sweden", "finland",
    "denmark", "norway", "portugal", "spain", "france",
    "amsterdam", "berlin", "dublin", "stockholm", "helsinki",
    "nl", "de", "ie", "se", "fi", "dk", "no", "pt", "es", "fr",
    "europe", "eu", "european"
  ],
  "require_visa_sponsorship": true,
  "label": "EU Visa Sponsored"
}

Targeting specific countries

Narrow allowed_regions to your target countries:

// Germany + Netherlands only
"allowed_regions": ["germany", "netherlands", "berlin", "amsterdam", "de", "nl"]

// Ireland + Portugal (English-friendly, popular for tech)
"allowed_regions": ["ireland", "portugal", "dublin", "lisbon", "ie", "pt"]

// Nordics
"allowed_regions": ["sweden", "finland", "norway", "denmark", "se", "fi", "no", "dk"]

What min_wfa_threshold means

ValueBehaviour
5Only triggers EU fallback when fewer than 5 global WFA jobs found
15 (recommended)Always triggers — ensures Adzuna AU/NZ/SG/UK always runs
50Effectively always-on regardless of WFA supply

Set to 15 or higher if you want Adzuna markets included every day.

Visa sponsorship reality check

  • Arbeitnow — filters for visa_sponsorship=true at the API level. Supply is genuinely thin (~10–50 jobs/day across all of Europe). Quality over quantity.
  • Adzuna AU/NZ/SG — no visa filter, but these markets have high English-language QA/AI/ML demand and straightforward work visa pathways compared to EU.
  • Adzuna UK — post-Brexit, UK has its own Skilled Worker visa. Strong tech market, no EU freedom of movement required.

Job boards by region

BoardRegionVisa filterNotes
ArbeitnowAll EU✅ YesBest for EU visa-sponsored roles
AdzunaAU, NZ, SG, UK❌ NoHigh volume, English-language
RemotiveGlobal WFA❌ NoStrong for fully remote
JobicyGlobal WFA❌ NoGood EU coverage
WeWorkRemotelyGlobal WFA❌ NoHigh quality listings

Project Structure

remote-job-hunter/
├── src/
│   ├── main.py          # Entry point, orchestration
│   ├── search.py        # 10+ platform scrapers + EU fallback
│   ├── scorer.py        # NLP resume scoring engine
│   ├── report.py        # Markdown report + WhatsApp message generation
│   ├── apply.py         # Auto-apply engine (Playwright)
│   ├── confirm.py       # Application confirmation workflow
│   ├── gaps.py          # Skill gap analysis
│   └── tailor.py        # Per-job resume tailoring
├── config/
│   ├── skill_keywords.json     # Domain keyword lists
│   └── profile.template.json  # Profile template
├── profiles/
│   ├── ai-ml.json              # AI/ML domain meta
│   └── qa-automation.json      # QA domain meta
├── .env.example
└── README.md

Roadmap

  • ✅ v1.0.0 — Multi-platform search, NLP scoring, skill gap analysis
  • ✅ v1.1.0 — WhatsApp delivery via OpenClaw cron
  • ✅ v1.2.0 — EU visa-sponsored fallback (Arbeitnow)
  • ✅ v1.3.0 — Adzuna AU/NZ/SG/UK integration, title exclude filter, scorer fixes
  • ✅ v1.3.1 — Resume pipe-expansion, QA keyword expansion, cron delivery fix
  • 🔄 v1.4.0 — Playwright auto-apply engine (libnss3 dependency fix + WSL support)
  • 📋 v1.5.0 — Cover letter generation per application
  • 📋 v1.6.0 — Application tracking dashboard
  • 📋 v2.0.0 — MCP server wrapper (integrate with MCPilot)

Requirements

requests
python-docx
pymupdf>=1.23.0
python-dotenv

Optional (for auto-apply):

playwright

Security

  • ✅ No credentials stored in profile/config files — all via .env
  • ✅ Outbound requests only to public job board APIs
  • ✅ All data written locally — nothing sent externally
  • ✅ Profile JSONs excluded from git via .gitignore

Author

Rajkiran Veldur — Senior AI/ML Solutions Architect
linkedin.com/in/rajkiranveldur · github.com/RajkiranVS/openclaw-remote-job-hunter


Built to solve my own job search problem. Now open source.

Ver o original no GitHub

Bom para

Relacionados