HR AI Assistant Skill

Purpose

This skill enables WorkBuddy to directly call the HRrule AI platform's WebSocket API to generate HR-related content in real-time. When users request HR policies, forms, consultation, professional Q&A, or other HR documents, this skill automatically invokes the API with appropriate parameters and streams the response back to the user. The skill supports:

1. 1. Document Generation: Creating HR policies, forms, job descriptions, performance reviews, etc.
2. Professional Q&A: Answering HR-related questions about labor laws, employee rights, compensation, contracts, etc.
3. Real-time Streaming: Providing immediate feedback with streaming responses

When to Use This Skill

Use this skill when users ask questions or make requests related to:

• - HR Policy Generation: Creating employee handbooks, attendance policies, leave policies, disciplinary procedures, compensation policies
• HR Form Creation: Generating onboarding forms, resignation forms, leave application forms, performance review forms
• HR Consultation: Asking questions about labor laws, employment contracts, recruitment processes, compensation and benefits
• HR Process Flowcharts: Creating visual representations of HR processes like recruitment workflows, onboarding processes
• HR Documentation: Writing reports, summaries, guidelines, and other HR-related documents
• Professional Q&A: Any HR-related questions requiring professional answers, such as labor law inquiries, employee rights, compensation, contract issues, leave policies, etc.

Key phrases that indicate this skill should be triggered include:

• - "帮我写一个员工手册"
• "生成HR制度"
• "制作入职表单"
• "HR相关问题"
• "劳动法咨询"
• "招聘流程"
• "薪酬制度"
• "离职手续"
• "考勤制度"
• "绩效考核"
• "员工福利"
• "劳动合同"
• "财务公司 需要招聘 社保专员，帮我生成一份 招聘JD"
• "三期员工有没有特殊保障？"
• "Generate HR policy"
• "Create HR form"
• "HR consultation"
• "Labor law questions"

How This Skill Works

Direct API Integration

This skill provides a Python script that interacts with the HRrule AI WebSocket API to generate content.

Integration Options:

1. 1. Analyze User Request: Parse the user's HR request to identify the content type
2. Check API Key: Verify API Key is configured (from env var or config.json)

- If API Key is missing: Display a friendly prompt to visit https://ai.hrrule.com/ for a free API Key - If API Key is valid: Proceed with API call

1. 3. Select Parameters: Automatically select appropriate tag_id and rt based on the request
2. Call API: The platform may execute the Python script to make the API call:

   python ~/.workbuddy/skills/hr-ai-assistant/scripts/call_hrrule_api.py \
       --content "用户的问题" \
       --tag-id 1 \
       --rt "专业问答"
   

1. 5. Stream Response: Stream the generated content back to the user in real-time
2. Handle Errors: Gracefully handle connection errors, timeouts, and API errors

Note: The platform should ensure safe script execution environments and validate all inputs.
Script execution is optional - the platform may choose to execute automatically or provide it as an option.

⚠️ Important: First-Time User Experience (首次使用体验)

When a user uses this skill for the first time without an API Key configured:

1. 1. Display Friendly Prompt:

CODEBLOCK1bash

运行配置脚本，交互式配置 API Key

python ~/.workbuddy/skills/hr-ai-assistant/scripts/configapikey.py
CODEBLOCK2

1. 2. Then proceed to answer using the platform's general AI model (WorkBuddy/OpenClaw's default model)

Example workflow:
CODEBLOCK3

NOTE: The prompt should be displayed BEFORE answering the question.

Tag ID and RT Parameter Mapping

The skill automatically maps user requests to the appropriate parameters:

Tag ID: 1 (专业问答 - Professional Q&A)

• - 通用 HR 咨询：劳动法问题、员工权益、薪酬福利、劳动合同、离职补偿、加班工资、试用期规定等任何 HR 相关的专业问答
• 示例问题：

- "三期员工有没有特殊保障？"
- "员工离职需要提前多少天通知？"
- "试用期工资可以低于转正工资多少？"
- "加班费怎么计算？"
- "员工旷工怎么处理？"
- "未签订劳动合同有什么后果？"
- "竞业限制补偿金标准是多少？"

Tag ID: 2 (制度类 - HR Policies)

• - 员工手册, 招聘管理制度, 入职试用期管理制度, 劳动合同管理制度, 薪酬管理制度, 考勤休假制度, 加班管理制度, 绩效管理制度, 员工培训制度, 离职管理制度, 员工竞聘制度, 病假管理制度, 奖惩制度, 其他制度

Tag ID: 7 (岗位类 - Job Descriptions)

• - 岗位说明书, 工作饱和度评估表, 任职资格标准, 职位图谱

Tag ID: 3 (绩效类 - Performance Management)

• - 360考核表, BSC考核表, KPI考核表, OKR考核表, 绩效承诺书, 绩效改进计划, 绩效面谈表, 绩效诊断报告

Tag ID: 4 (招聘类 - Recruitment)

• - 人才画像, 面试评估表, 面试题库, 招聘需求表, 招聘JD, 背景调查表, 录用条件说明书, 录用通知书, 入职承诺书, 劳动合同

Tag ID: 5 (薪酬类 - Compensation)

• - 薪酬等级表, 薪酬面谈表, 薪酬诊断报告, 岗位价值评估表

Tag ID: 8 (培训类 - Training)

• - 新员工培训计划, 年度培训计划

Tag ID: 13 (报告类 - Reports)

• - 年终总结, 月度报告, 周报, 日报

Tag ID: 14 (风控类 - Risk Management)

• - 风险自测

Usage Workflow

Step 1: Parse User Request

Analyze the user's request to determine:

• - Content Type: What type of HR content they need
• Tag ID: The appropriate content category
• RT: The specific resource type
• Details: Any specific requirements or constraints

Example Request Analysis:

Request: "财务公司 需要招聘 社保专员，帮我生成一份 招聘JD"

Analysis:

• - Content Type: 招聘JD (Job Description)
• Tag ID: 4 (招聘类)
• RT: "招聘JD"
• Details: 财务公司, 社保专员岗位

Step 2: Select Parameters

Based on the analysis, select the parameters:

CODEBLOCK4

Step 3: Call API Using Python Script

Use the bundled script to call the API:

CODEBLOCK5

Step 4: Stream Response to User

Stream the response back to the user in real-time:

CODEBLOCK6

Parameter Selection Logic

Automatic Mapping Rules

Rule 1: Keyword Matching

Keyword	Tag ID	RT
专业问答/咨询	1	专业问答
劳动法问题/咨询

1 | 专业问答 |
| 员工权益/权利 | 1 | 专业问答 |
| 离职补偿/赔偿 | 1 | 专业问答 |
| 加班工资/加班费 | 1 | 专业问答 |
| 试用期规定/工资 | 1 | 专业问答 |
| 三期/孕期/产期/哺乳期 | 1 | 专业问答 |
| 竞业限制 | 1 | 专业问答 |
| 经济补偿金/赔偿金 | 1 | 专业问答 |
| 违法解除/终止合同 | 1 | 专业问答 |
| 员工手册 | 2 | 员工手册 |
| 考勤制度 | 2 | 考勤休假制度 |
| 薪酬制度 | 2 | 薪酬管理制度 |
| 岗位说明书 | 7 | 岗位说明书 |
| KPI考核表 | 3 | KPI考核表 |
| 绩效考核表 | 3 | KPI考核表 |
| 绩效改进 | 3 | 绩效改进计划 |
| 面试评估表 | 4 | 面试评估表 |
| 面试题库 | 4 | 面试题库 |
| 招聘JD | 4 | 招聘JD |
| 招聘职位 | 4 | 招聘JD |
| 劳动合同 | 4 | 劳动合同 |
| 录用通知书 | 4 | 录用通知书 |
| 薪酬等级表 | 5 | 薪酬等级表 |
| 薪酬面谈表 | 5 | 薪酬面谈表 |
| 培训计划 | 8 | 新员工培训计划 |
| 年终总结 | 13 | 年终总结 |
| 周报/月报/日报 | 13 | 周报/月度报告/日报 |
| 风险评估 | 14 | 风险自测 |

Rule 2: Context Analysis

• - If user mentions "招聘" (recruitment) + specific document type → Tag ID: 4
• If user mentions "制度" (policy/system) → Tag ID: 2
• If user mentions "考核" (assessment) → Tag ID: 3
• If user mentions "培训" (training) → Tag ID: 8
• If user mentions "报告" (report) + time period → Tag ID: 13
• If user asks a question about labor laws, employee rights, compensation → Tag ID: 1 (专业问答)
• If user asks "有没有"、"如何"、"怎样"、"为什么"等疑问句 → Tag ID: 1 (专业问答)

Rule 3: Fallback - Professional Q&A

If no clear match with specific document types (如员工手册、招聘JD、KPI考核表等), use the general Q&A category:

• - Any HR-related question about laws, regulations, policies, employee rights → Tag ID: 1, RT: "专业问答"
• Example: "三期员工有没有特殊保障？", "员工离职需要提前多久通知？"
• Example: "试用期工资规定", "加班费计算方式"

Rule 4: Default Fallback

If still unclear, use the safest default:

• - HR policy/guidance → Tag ID: 2, RT: "其他制度"
• General HR question → Tag ID: 1, RT: "专业问答"

Script Usage

Script Location

IMPORTANT: The script is located at:
INLINECODE2

The platform should dynamically resolve this path using:
CODEBLOCK7

DO NOT use hardcoded paths like C:/Users/jikr/...

Function Signature

CODEBLOCK8

Parameters

• - api_key: API 密钥(必需)
• INLINECODE5: 用户问题或需求(必需)
• INLINECODE6: 内容类型 ID(必需)
• INLINECODE7: 资源类型(必需)
• INLINECODE8: WebSocket 地址(默认: ws://192.168.112.114:5000)
• INLINECODE9: 模型名称(默认: deepseek-ai/DeepSeek-R1)
• INLINECODE10: 收到内容块时的回调函数
• INLINECODE11: 完成时的回调函数
• INLINECODE12: 错误时的回调函数
• INLINECODE13: 是否显示详细日志

Return Value

Returns the complete response content as a string.

Command Line Usage

CODEBLOCK9

Example Workflows

Example 1: Generate Job Description (JD)

User Request:
CODEBLOCK10

Skill Execution:

1. 1. Parse Request:

- Identify: "招聘JD" → Tag ID: 4, RT: "招聘JD" - Extract context: 财务公司, 社保专员

1. 2. Construct Content:

CODEBLOCK11

1. 3. Call API:

CODEBLOCK12

1. 4. Stream Response:

The AI response is streamed back in real-time.

Example 2: Generate Employee Handbook

User Request:
CODEBLOCK13

Skill Execution:

1. 1. Parse Request:

- Identify: "员工手册" → Tag ID: 2, RT: "员工手册" - Extract context: 50人, 科技公司

1. 2. Call API:

CODEBLOCK14

Example 3: Create Performance Review Form

User Request:
CODEBLOCK15

Skill Execution:

1. 1. Parse Request:

- Identify: "KPI考核表" → Tag ID: 3, RT: "KPI考核表" - Extract context: 销售部门

1. 2. Call API:

CODEBLOCK16

Example 4: Professional Q&A (Labor Law Question)

User Request:
CODEBLOCK17

Skill Execution:

1. 1. Parse Request:

- Identify: Question about labor law/employee rights → Tag ID: 1, RT: "专业问答" - Extract context: 三期员工 (孕期、产期、哺乳期), 特殊保障

1. 2. Call API:

CODEBLOCK18

1. 3. Stream Response:

CODEBLOCK19

Example 5: General HR Consultation

User Request:
CODEBLOCK20

Skill Execution:

1. 1. Parse Request:

- Identify: Question about resignation → Tag ID: 1, RT: "专业问答" - Extract context: 员工离职, 提前通知时间

1. 2. Call API:

CODEBLOCK21

Error Handling

The script handles various error scenarios:

Connection Errors

• - Invalid WebSocket URL
• Network connectivity issues
• Firewall blocking

Authentication Errors

• - Invalid API Key
• API Key expired
• Permission denied

API Errors

• - Invalid parameters
• Server errors
• Timeout (120 seconds)

Error Handling Example

CODEBLOCK22

Best Practices

1. First-Time User Experience (首次使用体验)

Note: 当首次使用此 skill 且未配置 API Key 时，平台（WorkBuddy/OpenClaw）应该：

✅ 推荐做法：

1. 1. 检测到没有 API Key
2. 先显示友好的提示信息（访问 https://ai.hrrule.com/ 免费申请）
3. 然后使用平台的通用大模型回答问题
4. 用户手动配置 API Key（通过环境变量、配置文件或配置脚本）

应该显示的提示信息：
CODEBLOCK23

执行顺序：
CODEBLOCK24

注意：用户需要手动配置 API Key（通过环境变量或编辑配置文件），系统不会自动检测或保存。

2. Provide Detailed Context

Always include as much context as possible in the content parameter:

• - Company size and industry
• Specific requirements
• Target audience
• Time constraints

Good:
CODEBLOCK25

Less Good:
CODEBLOCK26

2. Use Streaming for Better UX

Always use the on_chunk callback to stream responses:

CODEBLOCK27

3. Handle Errors Gracefully

Always provide error callbacks:

CODEBLOCK28

4. Set Appropriate Timeout

The script has a 120-second timeout by default. For complex documents, this is usually sufficient.

Configuration

API Key Configuration

The API Key should be configured securely. Options:

1. 1. Environment variable: INLINECODE16
2. Configuration file
3. Direct parameter passing

WebSocket URL Configuration

Default: INLINECODE17

Can be changed via:

• - Command line: INLINECODE18
• Script parameter: INLINECODE19
• Configuration file

Troubleshooting

Issue: Connection Fails

Check:

• - WebSocket URL is correct
• Network connectivity
• Firewall settings
• Server is running

Issue: Authentication Error

Check:

• - API Key is valid
• API Key has not expired
• API Key has proper permissions

Issue: No Response

Check:

• - Content parameter is not empty
• Tag ID and RT are valid
• Network connection is stable
• Server is processing requests

Issue: Timeout

Solution:

• - Increase timeout in script
• Simplify the request
• Check server load

Integration Examples

DingTalk Integration

When receiving HR requests from DingTalk:

CODEBLOCK29

QQ Integration

Similar to DingTalk, parse the QQ message and call the API:

CODEBLOCK30

Summary

This skill provides direct API integration with the HRrule AI platform, enabling WorkBuddy to:

• - Automatically parse HR requests
• Select appropriate parameters (tag_id, rt)
• Call WebSocket API in real-time
• Stream responses back to users
• Handle errors gracefully

The skill is designed for seamless integration with chat platforms like DingTalk and QQ, making it easy to generate HR documents on demand.

Reference Materials

• - API Reference: INLINECODE20
• Example Prompts: INLINECODE21
• Usage Guide: INLINECODE22
• Test Page: INLINECODE23
• Original Test: INLINECODE24