skills.rest
zhansan379
zhansan379COMMUNITY
Explore all skills by this author →
WebsiteGitHub

boss-crawler

Crawl BOSS Zhipin job listings, match resumes with LLM scoring, and automate personalized applications.

89|6|Updated Aug 10, 2026
Search Tags:#resume matching#llm scoring#auto apply#boss zhipin#job crawler#drissionpage#job search automation
One-click install
npx skills add https://github.com/zhansan379/boss-crawler-skill --skill boss-crawler-zhansan379
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: boss-crawler
Source: https://github.com/zhansan379/boss-crawler-skill
Command: npx skills add https://github.com/zhansan379/boss-crawler-skill --skill boss-crawler-zhansan379

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires DrissionPage, requests, PyPDF2, python-docx, chardet, pillow, lxml, openpyxl, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve? Manually searching BOSS Zhipin, comparing each job against your resume, and writing individualized greetings for every HR contact is slow and repetitive. This Skill automates the full pipeline: crawl job postings, parse your resume, score matches with rules plus LLM analysis, generate tailored resumes and greetings, and deliver applications through browser automation. ## Core Features & Use Cases - Job Crawling & Matching: Collect listings via DrissionPage with keyword, city, and company-targeted modes, then score them with a fast rule engine or deep per-job LLM analysis. - Resume Intelligence: Parse PDF/Word/Markdown resumes into structured profiles, optimize content per JD without fabricating experience, and render privacy-safe resume images via the built-in ShowCV editor. - Automated Delivery: Send personalized greetings and resume images with rate limiting, dry-run preview, and explicit confirmation gates before any irreversible send. - Use Case: Upload your resume, let the Skill infer search parameters, crawl 50 matching positions, deep-match the top 20, review the HTML report, approve the shortlist, and auto-apply with tailored materials. ## Quick Start Ask Claude Code to run /boss-crawler and follow the prompts to upload your resume, crawl matching jobs, and deliver applications to the best matches.

Frequently Asked Questions about boss-crawler

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate job applications on BOSS Zhipin?▼

Run the pipeline stages in order: parse your resume, infer crawl parameters, crawl listings, match with quick or deep mode, generate greetings and optimized resumes, then execute apply.py with --yes. Delivery always requires explicit confirmation and includes a dry-run preview.

What LLM providers does this job matching tool support?▼

It defaults to the native Anthropic Messages protocol and also works with any OpenAI-compatible endpoint, including DeepSeek, Tongyi, Moonshot, Zhipu, Volcano Ark, and local ollama or vLLM. Configure base_url, api_key, and model in assets/llm_config.json, with per-stage model overrides.

Can I crawl jobs from specific companies only?▼

Yes, company-targeted mode (-m company) takes full company names and cities, fetches all their open positions via the search API with brandName filtering, and writes CSVs to assets/post_data/company/. These feed directly into the matching pipeline.

Does the resume optimization fabricate experience?▼

No. The verify stage cross-checks generated materials against the original resume text and flags technical terms with no source, blocking rendering when fabrication is detected. Availability fields like start date are copied only from the resume, never inferred.

Why does the crawler exit without collecting any jobs?▼

The most common cause is missing login state: the crawler exits normally when not logged in, so check for crawl_summary.json to confirm a real crawl. Run boss_post_interactive.py --ensure-login first to scan the QR code and persist the session.

What are the rate limits and safety measures for auto-applying?▼

Delivery enforces 3-5 second intervals between applications, caps sessions at 10-20 sends, pauses on captchas for manual handling, and requires an explicit --yes flag. Without --yes it only prints the planned recipient list.

Related Skills

Explore More Personal & Entrepreneur Skills→
lciaccilciacci

credentials

Discover, parse, validate, and configure API credentials for project environments.

Community
Intermediate
Imad-OuteImad-Oute
2

mh-cmo-follow-up-system

Schedule proposal follow-ups, draft messages, and capture win/loss outcomes.

Community
Intermediate
isaackehleisaackehle

job-search

Automate job research, scoring, resume tailoring, and interview preparation workflows.

Community
Advanced
HouseGarofaloHouseGarofalo
3

web-automation

Automate browser interactions and web scraping with Playwright, Puppeteer, Selenium, and Scrapy.

Official
Intermediate