can4hou6joeng4/boss-agent-cli

What it solves

boss-agent-cli is a command-line interface (CLI) designed to unify job discovery, application management, and recruitment processes across multiple platforms (like BOSS Zhipin and Zhilian). It solves the friction of manual job searching by providing a structured, AI-enhanced workflow for both job seekers and recruiters, allowing them to automate repetitive tasks like filtering for specific benefits, analyzing job descriptions, and managing candidate pools locally.

How it works

The tool operates as a schema-driven CLI that interacts with recruitment platforms via a browser bridge or API. It uses a "JSON envelope" for all standard output, making it easily consumable by both humans (via a terminal wizard) and AI agents (via MCP, subprocesses, or a Python SDK).

Key technical components include:

• Platform Abstraction: A registry system that allows the tool to work across different platforms (e.g., zhipin, zhilian).
• AI Integration: Supports OpenAI-compatible interfaces, Ollama, and vLLM for JD analysis, resume polishing, and interview preparation.
• Local Storage: Uses SQLite (WAL) for caching search results, managing a local shortlist of candidates, and storing encrypted login states.
• MCP Server: Exposes 73 tools to AI agents (like Claude Desktop or Cursor), enabling them to perform recruitment tasks autonomously.

Who it’s for

• Job Seekers: Those looking to automate job discovery, filter for specific benefits (e.g., "weekends off"), and use AI to optimize their resumes and cover letters.
• Recruiters: HR professionals who need a CLI-based approach to search for candidates, manage applications, and handle communications.
• AI Agent Developers: Developers building autonomous recruitment agents that need a structured, schema-driven interface to interact with job boards.

Highlights

• Benefit Filtering: Advanced filtering that performs real matching for specific benefits (e.g., "five insurances and one fund") across multiple pages.
• AI-Powered Enhancement: Built-in tools for JD analysis, resume optimization, and mock interviews using local or cloud LLMs.
• Robot-Agent Ready: Full MCP (Model Context Protocol) support, allowing AI agents to control the recruitment workflow.
• Dual-Role Workflow: Supports both candidate and recruiter roles with dedicated command sets for each.
• Resumable Crawling: A robust crawling system with checkpoints and budget constraints to prevent platform bans.
• JSON-First Output: Every command returns a structured JSON envelope, ensuring consistent integration with other tools.