← Back to Skills Marketplace
calmdowntr

Job Screener Engine

by CalmDownTR · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ Security Clean
79
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install job-screener-engine
Description
对用户提供的岗位机会进行结构化评估,输出多维量化评分和行动建议。
README (SKILL.md)

Job Screener Engine — 岗位筛选通用引擎

对用户提供的岗位机会进行结构化评估,输出量化评分和行动建议。 此版本为通用引擎,所有阈值和偏好均可通过 user_profile.md 自定义。


首次使用(必须执行)

在第一次运行之前,请引导用户完成配置:

  1. 检查 references/user_profile.TEMPLATE.md 是否存在
  2. 告诉用户需要填写以下信息:
    • Base 薪资底线、期望薪资
    • 核心技术方向(如 Agent/LLM / 前端 / 后端 / 嵌入式等)
    • 偏好的公司类型(如大厂 / 外企 / 国企 / 创业公司等)
    • 工作节奏要求
    • 目标城市
  3. 询问用户:"要不要我帮你用问答的方式生成 user_profile.md?"
    • 如果用户同意,根据 references/setup_wizard.md 逐一提问
    • 如果用户拒绝,告诉用户参考 TEMPLATE 自行填写
  4. 用户有 user_profile.md 后才执行正式评估

触发条件

当用户提供以下信息时自动触发:

  • 公司名称 + 岗位名称
  • JD 链接或截图内容
  • Boss 直聘等平台的岗位信息
  • "帮我看看这个岗位"、"值得投吗"等意图表达

工作流程

Step 1:解析用户输入

从用户消息中提取:

  • 公司名称
  • 岗位名称
  • JD 内容(如有)
  • 薪资区间(如有)
  • 其他用户已知信息

Step 2:信息收集

使用 web_search 补充以下信息:

  1. 公司基本面(必须搜索)

    • 搜索:[公司名] 融资[公司名] 上市
    • 目标:融资轮次、员工规模、成立时间
  2. 工作文化与口碑(必须搜索)

    • 搜索:[公司名] 工作体验 加班[公司名] 评价
    • 目标:加班情况、双休、管理评价
  3. 行业信号(如有时间则搜索)

    • 搜索:[公司名] 最新消息[公司名] 新闻
    • 目标:近期重大新闻、行业趋势

Step 3:一票否决预检

读取 references/user_profile.md 获取用户的一票否决条件,检查是否触发。 若触发则直接判定 ❌,跳到 Step 6 输出报告(总分栏写「一票否决」)。

默认一票否决条件(用户可在 user_profile.md 中覆盖):

  • Base 低于用户底线且无谈判空间
  • 明确 996 或严重内卷文化
  • 公司有严重负面新闻
  • 岗位与用户技术方向完全无关

Step 4:评分

读取 references/scoring_framework.md 获取评分框架和标准。 同时读取 references/user_profile.md 获取用户的自定义权重和阈值。

按六个维度打分(满分100):

维度 默认权重 说明
薪资匹配度 18分 对比用户 Base 底线
公司规模与规范度 20分 融资/规模/管理
技术成长空间 15分 与用户技术方向的匹配
工作节奏与心理适配度 15分 加班/管理风格
稳定性 20分 资金来源判断
区域居住可行性 12分 区域生活适配

评分规则:

  • 有明确信息的维度:按标准打分
  • 信息不足的维度:按中间值处理,并在输出中标注「信息不足」

Step 5:信息不足处理

如果关键维度(薪资、公司规模、工作节奏)信息不足:

  • 列出需要用户补充的具体问题(最多3个)
  • 使用简洁格式提问
  • 用户无法补充则按中间值处理

Step 6:输出评估报告

使用以下格式输出:

## 📋 岗位评估 | [公司名] - [岗位名]

### 基本信息
- **薪资:** [区间/面议]
- **公司规模:** [轮次/人数/上市情况]
- **地点:** [城市]
- **信息来源:** [用户提供的 + 搜索补充的]

### 评分明细

| 维度 | 分值 | 说明 |
|------|------|------|
| 薪资匹配度 | X/18 | [一句话理由] |
| 公司规模 | X/20 | [一句话理由] |
| 技术成长 | X/15 | [一句话理由] |
| 工作节奏 | X/15 | [一句话理由] |
| 稳定性 | X/20 | [一句话理由] |
| 区域可行性 | X/12 | [一句话理由] |
| **总分** | **X/100** | |

### 判定:[⭐强烈推荐 / ✅推荐投递 / ⚠️可以考虑 / 🔶勉强可投 / ❌不建议]

### 重要程度:[S / A / B / C]

### 行动建议
[1-3条具体建议]

### 面试练手建议
[高价值 / 中等价值 / 低价值]

**高价值:** 岗位与用户技术方向相关,公司愿意给面试,即使不会接受 offer 也值得练手。
**中等价值:** 岗位部分相关,或薪资/稳定性有明显短板,但面试经验仍可积累。
**低价值:** 岗位完全不相关,或公司有严重问题(996/诈骗/传销),不值得花时间。

> **注意:** "❌不建议入职" ≠ "不值得面试"。本字段独立评估面试练手价值。

### ⚠️ 风险提示(如有)
[列出需要面试时重点验证的问题]

Step 7:同分排序(多个岗位对比时)

当两个岗位总分接近(差值≤5分)时,参照 references/user_profile.md 中的"同分排序规则"进行优先级排序。


注意事项

  • 不替用户做最终决定,只提供量化依据和建议
  • 薪资以 Base 为准,年终奖/期权不作为达标条件
  • 信息不足时诚实标注,不猜测不编造
  • 多个岗位同时评估时,输出对比表格便于用户排序
  • 保持客观,不因用户对某公司有好感而抬高分数
  • 用户未配置 user_profile.md 时:先执行配置向导,不评估

参考资料

  • references/scoring_framework.md — 评分框架与维度标准
  • references/setup_wizard.md — 首次配置引导
  • references/info_checklist.md — 信息收集清单
  • references/user_profile.md — 用户画像(需用户自行填写)
  • references/user_profile.TEMPLATE.md — 用户画像模板
Usage Guidance
This skill appears safe to install if you are comfortable creating a local profile with your salary and job preferences and allowing the agent to search the web for company information. Keep the generated profile private and do not include details you would not want stored locally.
Capability Analysis
Type: OpenClaw Skill Name: job-screener-engine Version: 1.0.0 The job-screener-engine is a structured tool designed to help users evaluate job opportunities by combining user-provided data with web search results. It includes a setup wizard to create a local user profile containing sensitive information (salary expectations, tech stack, and location), which is necessary for its stated purpose of quantitative job scoring. The analysis found no evidence of malicious intent, data exfiltration, or unauthorized command execution; the use of 'WebSearch' and 'Read' tools is consistent with the documented workflow in SKILL.md and references/info_checklist.md.
Capability Assessment
Purpose & Capability
The skill's purpose, reading its own reference files, and using WebSearch to research companies all fit the stated job-evaluation function.
Instruction Scope
Instructions are limited to parsing job information, collecting missing details, scoring roles, and producing recommendations; they do not attempt to override user intent or force unrelated behavior.
Install Mechanism
No install spec or code files are present, and the artifacts describe an instruction-only skill with no required binaries, environment variables, or credentials.
Credentials
The skill expects a local user_profile.md containing salary and career preferences and uses web search for company reputation checks; both are proportionate but worth user awareness.
Persistence & Privilege
The skill creates and later reads a persistent local profile file, but the file path is scoped to the skill's references directory and no elevated privileges or credentials are requested.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install job-screener-engine
  3. After installation, invoke the skill by name or use /job-screener-engine
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Job Screener Engine v1.0.0 - Initial release of a structured job evaluation engine for scoring and advice. - Supports user-customized preferences and thresholds via `user_profile.md`. - Guides first-time users through a setup process for personalization. - Automates information gathering using web search for company data, job reviews, and news. - Applies veto rules and six-dimensional scoring to job opportunities. - Outputs detailed, clearly formatted reports with actionable suggestions.
Metadata
Slug job-screener-engine
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is Job Screener Engine?

对用户提供的岗位机会进行结构化评估,输出多维量化评分和行动建议。 It is an AI Agent Skill for Claude Code / OpenClaw, with 79 downloads so far.

How do I install Job Screener Engine?

Run "/install job-screener-engine" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Job Screener Engine free?

Yes, Job Screener Engine is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Job Screener Engine support?

Job Screener Engine is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Job Screener Engine?

It is built and maintained by CalmDownTR (@calmdowntr); the current version is v1.0.0.

💬 Comments