OpenClaw 职位搜索
一个为 OpenClaw 构建的结构化职位搜索工作流，作为 薄型本地插件加仓库拥有的技能。
项目是什么
该项目分为两层：
这种分离保持了插件的轻量和易安装性，同时允许工作流使用 Python 和通过技能的更丰富的操作脚本。
帮助您做什么
- 准备有组织的职位搜索运行
- 通过 JobSpy 和技能脚本检索列表
- 使用 OpenClaw 代理评估列表
- 生成定制的 JSON 简历文件
- 渲染简历为 PDF
- 将审查结果导出到 Excel
主要技能
环境检查
首先使用此技能，当设置或运行时准备不明确时。
检查：
- Python 可用性
- JobSpy 依赖项
- Node 依赖项
- Puppeteer 可用性
- 简历主题解析
职位搜索技能
拥有的功能：
- 候选人理解
- 查询规划
- 运行准备
- 检索过滤器
- 通过技能脚本执行 Python 检索
职位列表评估技能
拥有的功能：
- 单列表评估逻辑
- 保留 / 可能 / 删除决策
- 规范化评分输出
- 确定性评估工件写入
职位简历生成技能
拥有的功能：
- 真实的适度定制
- 为选定的列表生成 JSON 简历
- 为选定的简历渲染 PDF
捆绑技能拥有的能力
此包现在是显式的技能优先。
插件通过 openclaw.plugin.json 捆绑其技能，当启用插件时，OpenClaw 应该将其加载为捆绑插件技能。
确定性操作位于每个技能的 scripts/ 文件夹中，而不是在插件入口处。
典型工作流
- 如果设置不确定，运行环境检查技能。
- 使用职位搜索技能：
- 读取个人资料
- 准备查询
- 通过技能拥有的脚本准备运行
- 通过 Python 脚本执行检索
- 使用评估技能评估列表。
- 使用简历生成技能生成定制的 JSON 简历文件。
- 作为简历工作流的一部分，渲染选定的简历为 PDF。
- 通过技能拥有的脚本写入评估和导出工件。
安装
openclaw plugins install @emrekuruu/job-search openclaw plugins enable job-search
然后在仓库内：
为什么这样结构化
项目故意将操作逻辑移到捆绑的技能和技能脚本/引用中，留下了最薄的本地插件表面。
这给您：
- 更干净的本地插件表面
- 更容易的技能迭代
- 无需将进程启动逻辑塞入插件层的 Python 检索
Job Search for OpenClaw
A structured OpenClaw job-search workflow built as a thin native plugin plus repo-owned skills.
What this project is
This project is split into two layers:
- Minimal plugin entrypoint
- Skills
That split keeps the plugin small and install-friendly while letting the workflow still use Python and richer operational scripts through skills.
What it helps with
- prepare organized job-search runs
- retrieve listings with JobSpy through skill scripts
- evaluate listings with OpenClaw agents
- generate tailored JSON Resume files
- render resumes to PDF
- export reviewed results to Excel
Main skills
environment-check
Use this first when setup or runtime readiness is unclear.
Checks:
- Python availability
- JobSpy dependencies
- Node dependencies
- Puppeteer availability
- resume theme resolution
job-search-skill
Owns:
- candidate understanding
- query planning
- run preparation
- retrieval filters
- Python retrieval execution through skill scripts
job-listing-evaluation-skill
Owns:
- one-listing evaluation logic
- keep / maybe / drop decisions
- normalized scoring output
- deterministic evaluation artifact writing
job-resume-generation-skill
Owns:
- truthful moderate tailoring
- JSON Resume generation for selected listings
- PDF-render step for selected resumes
Bundled skills own the capabilities
This package is now explicitly skills-first.
The plugin bundles its skills through openclaw.plugin.json, and OpenClaw should load them as bundled plugin skills when the plugin is enabled.
Deterministic operations live under each skill's scripts/ folder rather than in the plugin entrypoint.
Typical workflow
- Run the environment check skill if setup is uncertain.
- Use the job search skill to:
- read the profile
- prepare queries
- prepare the run via the skill-owned script
- execute retrieval via the Python script
- Evaluate listings with the evaluation skill.
- Generate tailored JSON Resume files with the resume-generation skill.
- As part of that resume workflow, render selected resumes to PDF.
- Write evaluation and export artifacts through the skill-owned scripts.
Install
openclaw plugins install @emrekuruu/job-search
openclaw plugins enable job-search
Then inside the repo:
Why it is structured this way
The project intentionally moves operational logic into bundled skills and skill scripts/references, leaving the plugin index as thin as possible.
This gives you:
- a cleaner native plugin surface
- easier skill iteration
- Python retrieval without stuffing process-launch logic into the plugin layer