convez.ai
AgenteAutomaçãoAvançado
dpgit07

Job Search Automation Agent

dpgit07/job-search-agent

Automação completa de busca de emprego com IA — scan, tailor, track, prep

Ver no GitHub
InicialPluginScore · de 0 a 100

Agente de busca de emprego alimentado por IA que automatiza varredura de vagas, personalização de currículos, análise de gaps de habilidades e preparação para entrevistas. Integra-se com LinkedIn, Indeed e páginas de carreira, mantendo rastreamento estruturado de candidaturas e gerando materiais customizados para cada oportunidade.

Instalação

$npx convez-ai dpgit07/job-search-agent

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
0
Forks
0
Acompanhando
0
Saúde
28%
Último commit
há 1 dia
Versão
—
Issues abertas
0
Contribuidores
2

Casos de uso

  • Automatizar varredura diária de vagas em múltiplas plataformas
  • Gerar currículos e cartas de apresentação personalizados por vaga
  • Identificar gaps de habilidades e priorizar upskilling
  • Preparar materiais técnicos e comportamentais para entrevistas
  • Manter registro centralizado de candidaturas e follow-ups

Veredito

Agente robusto e bem-arquitetado para quem busca profissionalizar a jornada de candidatura. Usa subagents especializados, integração com MCPs (Apify, Gmail, Sheets) e orquestração via Claude Code — excelente para papéis técnicos sênior. Requer setup inicial cuidadoso, mas o ROI em tempo economizado é alto.

README

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

🤖 Job Search Automation Agent

An AI-powered job search agent built on Claude Code — automates scanning, resume tailoring, cover letter writing, interview prep, and application tracking for Senior Data Engineer roles in the USA.

Author: Dharmendra Kumar Prajapati
Stack: Claude Code · Apify MCP · Claude Subagents · Slash Commands · Hooks


What This Does

ModuleWhat It Automates
Job ScannerScrapes LinkedIn, Indeed & company career pages daily for matching roles
Resume TailorGenerates ATS-optimized resume variants per job description
Skill Gap AnalyzerCompares your profile vs. market trends, ranks upskilling priorities
Application TrackerMaintains a structured JSON tracker of all applications & follow-ups
Interview PrepGenerates company-specific technical + behavioral prep packages
Networking AssistantDrafts personalized recruiter outreach, referral requests, thank-you notes
Cover Letter WriterWrites tailored, ATS-friendly cover letters (never generic)

Architecture

┌─────────────────────────────────────────────────────────┐
│                CLAUDE CODE (Main Agent)                  │
│                      CLAUDE.md                           │
│           (Profile · Rules · Preferences)                │
├──────────────┬──────────────┬────────────┬──────────────┤
│  Subagent:   │  Subagent:   │ Subagent:  │  Subagent:   │
│  job-scanner │resume-tailor │ skill-gap  │interview-prep│
├──────────────┴──────────────┴────────────┴──────────────┤
│                     MCP SERVERS                          │
│  Apify (scrape) · Gmail · Google Sheets · GitHub         │
│  Google Drive · Google Calendar · Filesystem             │
├─────────────────────────────────────────────────────────┤
│  HOOKS: SessionStart → load state · Stop → save memory  │
└─────────────────────────────────────────────────────────┘

Project Structure

job-search-agent/
├── CLAUDE.md                        # Core agent instructions & candidate profile
├── README.md
├── .gitignore
│
├── resume/
│   ├── master_resume.md             # Your base resume (edit this)
│   └── tailored/                    # Auto-generated tailored versions
│
├── data/
│   ├── profile.json                 # Your skills, experience, preferences
│   ├── job_tracker.json             # Application database (auto-updated)
│   ├── skill_market_trends.json     # Cached market analysis
│   ├── target_companies.json        # Tier 1–4 target companies
│   └── interview_notes/             # Company-specific prep materials
│
├── output/
│   ├── reports/                     # Daily/weekly digest reports
│   ├── cover_letters/               # Generated cover letters
│   └── tailored_resumes/            # Final tailored resume files
│
└── .claude/
    ├── settings.json                # Permissions, hooks config
    ├── memory.md                    # Persistent session memory
    ├── agents/                      # 6 subagent definitions
    │   ├── job-scanner.md
    │   ├── resume-tailor.md
    │   ├── skill-gap-analyzer.md
    │   ├── interview-prep.md
    │   ├── networking-assistant.md
    │   └── cover-letter-writer.md
    └── commands/                    # 8 slash commands
        ├── scan.md       → /scan
        ├── apply.md      → /apply
        ├── tailor.md     → /tailor
        ├── prep.md       → /prep
        ├── status.md     → /status
        ├── trends.md     → /trends
        ├── outreach.md   → /outreach
        └── daily.md      → /daily

Slash Commands

CommandUsageWhat It Does
/daily/dailyFull morning routine: scan → tailor → status → digest
/scan/scanFind new matching jobs (70%+ fit) across all sources
/apply/apply GoogleTailor resume + cover letter for a job, then confirm to submit
/tailor/tailor AmazonGenerate ATS-optimized tailored resume for a JD
/prep/prep MetaGenerate interview prep package (technical + behavioral)
/status/statusFull application dashboard with follow-up alerts
/trends/trendsMarket skill demand analysis + your gap report
/outreach/outreach recruiter DatabricksDraft personalized networking message

Setup

Prerequisites

  • Claude Code installed (npm install -g @anthropic-ai/claude-code)
  • Node.js 18+ and Python 3.10+
  • An Apify account (free tier: 100 scrapes/month)
  • A GitHub account

Step 1 — Clone & Enter the Project

git clone https://github.com/dpgit07/job-search-agent.git
cd job-search-agent

Step 2 — Update Your Profile

Edit these two files with your personal details:

CLAUDE.md — update:

- Visa Status: [H1B / GC / Citizen]
- LinkedIn: [your LinkedIn URL]

data/profile.json — update:

"visa_status": "H1B",
"linkedin_url": "https://linkedin.com/in/yourprofile"

resume/master_resume.md — paste in your full resume content.

Step 3 — Install MCP Servers

# Minimum required — job scraping
claude mcp add apify \
  -e APIFY_TOKEN=your_apify_token \
  -- npx -y @anthropic-ai/apify-mcp-server

# Local file access
claude mcp add filesystem \
  -- npx -y @modelcontextprotocol/server-filesystem $(pwd)

# Optional — GitHub version control
claude mcp add github \
  -e GITHUB_PERSONAL_ACCESS_TOKEN=your_github_token \
  -- npx -y @modelcontextprotocol/server-github

# Optional — Gmail, Calendar, Sheets (requires Google OAuth setup)
claude mcp add gmail -- npx -y @anthropic-ai/gmail-mcp-server
claude mcp add google-sheets -- npx -y @anthropic-ai/google-sheets-mcp-server
claude mcp add google-calendar -- npx -y @anthropic-ai/google-calendar-mcp-server

Get your Apify token from: https://console.apify.com/account/integrations

Step 4 — Launch

claude
# Inside Claude Code:
/daily

Daily Workflow

Morning:
  /daily              → Full scan + digest + action list

Per application:
  /apply {company}    → Tailored resume + cover letter (requires your OK before submitting)
  /outreach recruiter {company}  → Draft recruiter message

Before interviews:
  /prep {company}     → Technical + behavioral prep package

Weekly:
  /trends             → Skill gap report + market intelligence
  /status             → Full pipeline dashboard

Fit Scoring

Each job is scored 0–100 using weighted criteria:

CriterionWeight
Technical skill match30%
Experience level match20%
Salary / compensation15%
Company tier (1–4)10%
Location / remote flexibility10%
Growth opportunity10%
Visa compatibility5%

Jobs scoring ≥ 85% are flagged HIGH PRIORITY and get automatic resume/cover letter drafts.
Jobs scoring < 70% are filtered out.


Safety & Privacy

  • Never auto-submits — Every application requires explicit yes confirmation
  • Data stays local — Profile and tracker live in data/ on your machine
  • No secrets in repo — .gitignore excludes all tokens, credentials, and OAuth files
  • Scraping via Apify — Uses Apify's authorized infrastructure (rate-limit safe)

Target Companies

TierCompanies
Tier 1LinkedIn, Meta, Google, Apple, Amazon, Microsoft, Netflix
Tier 2Databricks, Snowflake, Capital One, Airbnb, Uber, Stripe, Disney
Tier 3T-Mobile, Salesforce, Adobe, Oracle, Walmart, JPMorgan, Goldman Sachs
Tier 4Any company · matching role · $140K+ base · good culture

MCP Servers Reference

ServerPurposeRequired?
apifyJob scraping from LinkedIn, Indeed, company sitesYes
filesystemRead/write local project filesYes
githubResume version controlRecommended
gmailEmail applications, track responsesOptional
google-sheetsSpreadsheet application trackerOptional
google-calendarSchedule interviews, set remindersOptional

Contributing

This is a personal job search automation tool. Fork it and adapt the CLAUDE.md candidate profile section to your own background.


Built with Claude Code · Powered by Apify · v1.0.0

Ver o original no GitHub

Bom para

Relacionados