liepin-workbuddy-mcp
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@liepin-workbuddy-mcp 用猎聘搜索深圳的产品经理岗位,期望薪资 25k-35k"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
liepin-workbuddy-mcp
一个供 WorkBuddy 使用的本地 stdio MCP Server 适配层,把猎聘官方命令行工具 liepin-cli 的能力(职位搜索、简历读取、求职期望更新、职位投递)暴露为 MCP 工具。
本项目不自行实现猎聘 API 客户端——不调用未公开接口、不抓取页面、不模拟登录,只通过官方授权页链接的 liepin-cli 完成操作。
⚠️ 第三方说明:官方
liepin-cli仓库目前未声明开源许可证,因此本仓库不 vendoring、不复制、不再分发其源码,只依赖用户本机已安装的 CLI 命令。猎聘 CLI 的安装与授权请以官方授权页为准。
能力一览
MCP 工具 | 作用 | 是否写操作 |
| 按岗位、城市、薪资、公司等条件搜索职位 | 只读 |
| 读取当前账号简历 | 只读 |
| 更新求职期望 | 写(需明确确认) |
| 投递指定职位 | 写(需明确确认,逐职位核对) |
投递与更新都是不可轻易撤回的动作:适配器为两者设置了精确的确认文本门槛,缺失确认会直接阻止调用,杜绝批量或猜测式投递。
Related MCP server: MCP Server for LinkedIn
项目结构
yx/
├── connector-meta.json # WorkBuddy 连接器元信息
├── mcp.json # MCP 配置样例(按本机修改 command 路径后使用)
├── pyproject.toml
├── src/liepin_workbuddy_mcp/
│ ├── __init__.py
│ ├── cli.py # 对 liepin-cli 的安全子进程封装
│ └── server.py # FastMCP stdio server
├── skills/liepin-job-search/SKILL.md
└── tests/安装与配置
1. 安装官方猎聘 CLI
猎聘 CLI 安装与授权遵循官方授权页:https://www.liepin.com/mcp/auth(授权页所链接仓库为 https://github.com/liepin-tech-2026/liepin-cil)。任选官方 README 中的一种方式安装,确保 liepin-cli 在 PATH 中可调用。
首次授权在可交互终端中执行:
liepin-cli setup按提示打开授权页、登录,把页面生成的 Token 粘贴到终端。Token 只保存在本机 CLI 配置中,不要发到聊天、截图或项目文件;凭证有效期 90 天。
检查授权状态:
liepin-cli auth status2. 安装本项目
python -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"3. 在 WorkBuddy 中配置 MCP
项目根目录的 mcp.json 是配置样例。将其中的 mcpServers.liepin-workbuddy 节点合并到 WorkBuddy 的用户级 MCP 配置(~/.workbuddy/mcp.json),并按本机实际安装位置修改 command 路径;不要覆盖已有的其它 MCP server 项。合并后在 WorkBuddy 连接器管理页面找到 liepin-workbuddy,点击「信任」并启用。
提示:MCP server 通过 stdio 通信,进程本身不打印启动横幅或调试日志。
4. 开始使用
「用猎聘搜索深圳的产品经理岗位,期望薪资 25k–35k。」
「读取我的猎聘简历,帮我总结优势和需要补强的地方。」
「搜索符合条件的岗位,逐个比较;投递前先把岗位信息给我确认。」
薪资单位、薪资类型等筛选值以 liepin-cli job search --help 展示的格式为准;不确定时先查帮助,不要猜测编码。职位返回为 CLI 的原始 JSON,本项目透传、不承诺字段形状。
本地运行与测试
.venv/bin/python -m pytest -q也可在终端直接启动 MCP server:
.venv/bin/liepin-workbuddy-mcp安全边界与限制
写操作(更新求职期望、投递)必须在用户查看字段并明确确认后才执行。
职位搜索结果中的
job_id、job_kind均须原样取自本次结果,禁止推断或捏造。简历数据涉及个人信息,仅在当前授权账号与用户确认的范围内调用,不另行保存副本。
官方 CLI 未声明许可证,本项目不随附其源码;因此无真实授权账号时,集成测试无法验证线上数据,单元测试使用假 CLI 验证命令组装、JSON 解析、错误脱敏与确认门槛。
Available Tools
4 toolsliepin_apply_jobA
向猎聘投递指定职位。投递不可轻易撤回,必须先展示职位详情并取得用户明确确认。
job_id 和 job_kind 必须来自本次猎聘职位搜索结果;job_kind 不得猜测。confirmation 必须精确为“确认投递该猎聘职位”。
Inputs: job_id (required), job_kind (required), confirmation (required). Returns: result
liepin_get_resumeA
读取当前猎聘账号简历。仅返回猎聘官方 CLI 的原始 JSON 数据,不会保存副本。
Inputs: no named parameters. Returns: result
liepin_search_jobsB
使用猎聘官方 CLI 搜索职位。薪资和筛选值按猎聘 CLI 支持格式原样传入;page 从 0 开始。
返回官方 CLI 原始 JSON,不假设字段结构。调用前应确认目标城市、岗位和薪资条件已明确。
Inputs: page, address, job_name, edu_level, +6 more. Returns: result
liepin_update_job_expectationA
更新猎聘的求职期望。写操作必须在用户查看更新字段后获得其明确确认。
confirmation 必须精确为“确认更新猎聘求职期望”;record_id 必须来自简历查询结果。 支持 jobtitle、dq、wantSalaryLow、wantSalaryHigh、wantSalaryMonths、workType、workweek 和 practiceMonths。
Inputs: record_id (required), dq, jobtitle, workType, +6 more. Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 4 tools
Each tool maps to a distinct action in the job-search workflow: search jobs, read resume, update job expectations, and apply to a job. There is no meaningful overlap between their purposes, and the write-operation confirmations further distinguish them.
All tool names follow a clear and consistent liepin_verb_noun snake_case pattern. The verbs (search, get, update, apply) accurately describe each operation.
Four tools is a well-scoped set for a focused job-search assistant, and every tool has a clear role. The count is neither bloated nor trivially thin.
The surface covers search, resume reading, expectation updates, and application submission, but notable lifecycle gaps remain. There is no dedicated job-detail retrieval, application status tracking, withdrawal, or resume update operation, which may limit agents after submission.
Related MCP Connectors
Liepin job search and resume workflows backed by the official Liepin MCP server.
A hosted LinkedIn MCP server for your own account. Paste one address into Claude, ChatGPT, Cursor or n8n to read your inbox, send messages and invitations, and search people. Free to start. Not made by LinkedIn.
Talk to your public-facing AI from any MCP client — Claude, ChatGPT, Cursor, Cline, Windsurf.
Managed LinkedIn MCP server for AI agents: search, connect, message and enrich on accounts you own.
Related MCP Servers
- AlicenseBqualityAmaintenanceLocal MCP server for BOSS Zhipin workflows. Exposes 49 tools for job search, welfare filtering, recruiter messaging, pipeline tracking, and resume optimization for AI agents.7910,905 PyPI2,224MIT
- AlicenseAqualityDmaintenanceAn MCP server that lets AI assistants like Claude read LinkedIn data through your own logged-in browser session. Access profiles and companies, search for jobs, or get job details.19Apache 2.0
- AlicenseNot gradedqualityCmaintenanceEnables any MCP host to search job postings on sinojobs.com and retrieve individual job details via read-only search and lookup tools, with keyword, location, remote-only, and posting-date filters. Works over stdio with polite rate limiting and returns normalized results including title, company, location, apply URL, and publish date.MIT
- AlicenseNot gradedqualityCmaintenanceEnables AI agents to search live job listings on waiqipin.cn and fetch individual roles with application links, filtering softly by keyword, location, recency, and remote suitability. Runs locally over stdio so compatible MCP hosts gain read-only access to those public opportunities.MIT