Contract Risk Review

io.github.justingyang218Published by GitHub account justingyang218v1.0.0Updated Oct 24, 2026

Chinese labor & service contract risk assessment with 25 categories and local law differences.

The publisher's own server: Review Expert

VerifiedSTDIODesktop onlyKnowledge & MemoryBusiness & Commerce

Overview

AI-generated overview

A local MCP server that reviews Chinese labor and service contracts for risk, citing 25 risk categories and regional legal differences.

What you can do
Review a single labor or service contract for risk, with automatic contract-type detection and masking of ID numbers, phone numbers and bank card numbers in the output. Batch-review several contracts at once and produce an overview table. Look up risk-review points and legal basis for a given risk category. Look up regional regulation differences for Beijing, Shanghai, Guangdong, Zhejiang and Jiangsu on topics such as non-compete compensation, overtime pay base, minimum wage, maternity leave and marriage leave. Browse risk-category lists and region lists as resources, and use built-in review and batch-review prompt templates.
When to use it
Useful when you need a first-pass risk check on Chinese employment or labor-service contracts, want clause-level review points and legal references, or need to compare how a topic is treated across the supported provinces and cities. It is a reference aid, not a substitute for a licensed lawyer.
Requirements
Runs as a local process (Python, installed from requirements; also offered as an MCP bundle). Requires LLM_API_KEY, plus LLM_BASE_URL and LLM_MODEL for an OpenAI-compatible endpoint; LLM_TEMPERATURE is optional. Network access to that LLM endpoint is needed. Can run over stdio or as an HTTP service on a local port.
Before you install
Contract text is sent to the configured OpenAI-compatible LLM endpoint, so choose a provider you trust with the content. LLM_API_KEY is a secret and should be supplied through environment configuration, not hard-coded. The server states its output is only a risk reference and does not constitute formal legal advice or replace a practicing lawyer's review.

Installation

In SourceWeft

  1. Open Contract Risk Review in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

劳动&劳务合同风险排查 MCP Server

基于 FastMCP 构建的合同风险审查 MCP 服务，可被任何支持 MCP 协议的客户端（DeepSeek Harness、Claude Code、Cursor、Windsurf 等）调用。

✨ 功能特性

  • 单份合同审查：对劳动合同（16类风险）或劳务合同（9类风险）进行全面智能审查
  • 批量合同审查：支持多份合同批量审查，自动生成总览表
  • 风险知识库查询：查询指定类别的风险审查要点、法律依据
  • 地方法规查询：查询北京/上海/广东/浙江/江苏的地方法规差异
  • 自动合同类型识别：智能判断是劳动合同还是劳务合同
  • 个人信息脱敏：自动对身份证号、手机号、银行卡号进行脱敏处理

📦 安装

bash
pip install -r requirements.txt

⚙️ 配置

环境变量

变量名必填默认值说明
LLM_API_KEY✅ 是-大模型 API 密钥
LLM_BASE_URL❌ 否https://api.openai.com/v1API 端点地址（OpenAI 兼容接口）
LLM_MODEL❌ 否gpt-4o-mini模型名称
LLM_TEMPERATURE❌ 否0.1采样温度

配置示例

bash
export LLM_API_KEY="sk-xxxxxxxxxxxxxx"export LLM_BASE_URL="https://api.deepseek.com/v1"export LLM_MODEL="deepseek-chat"

💡 支持任何 OpenAI 兼容的 API 接口（DeepSeek、通义千问、智谱 AI、Ollama 等）

🚀 运行

方式一：STDIO 模式（本地工具/桌面客户端）

bash
python server.py# 或显式指定python server.py stdio

方式二：HTTP 模式（网络服务）

bash
python server.py http 0.0.0.0 8000

启动后访问 http://localhost:8000/mcp 即可使用。

方式三：使用 fastmcp CLI

bash
# 开发调试fastmcp dev server.py
# 安装到 Claude Desktopfastmcp install server.py

🔌 MCP Tools

1. review_contract

对合同文本进行完整的风险审查。

参数：

  • contract_text (string, 必填)：合同文本内容
  • contract_type (string, 可选)：合同类型，auto/labor/service，默认 auto

示例：

review_contract(contract_text="劳动合同文本...", contract_type="auto")

2. batch_review

批量审查多份合同。

参数：

  • contracts_text (string, 必填)：多份合同文本
  • separator (string, 可选)：分隔符，默认 ---合同分割线---

示例：

batch_review(contracts_text="合同1\n---合同分割线---\n合同2", separator="---合同分割线---")

3. get_risk_knowledge

查询指定类别的风险知识库。

参数：

  • category (string, 必填)：风险类别，如"试用期"、"竞业限制"、"人身损害赔偿"

示例：

get_risk_knowledge(category="竞业限制")

4. get_local_regulations

查询指定地区的法规差异。

参数：

  • region (string, 必填)：地区，北京/上海/广东/浙江/江苏
  • topic (string, 必填)：主题，如竞业补偿/加班费基数/最低工资/产假/婚假

示例：

get_local_regulations(region="上海", topic="竞业补偿")

📚 MCP Resources

URI说明
risk-knowledge://categories所有风险类别清单
risk-knowledge://labor劳动关系16类风险清单
risk-knowledge://service劳务关系9类风险清单
local-regions://list支持的地区列表

💬 MCP Prompts

名称说明
contract_review_prompt完整合同审查提示模板
batch_review_prompt批量审查提示模板

🔗 集成到客户端

Claude Desktop

在 claude_desktop_config.json 中添加：

json
{  "mcpServers": {    "contract-risk-review": {      "command": "python",      "args": ["/path/to/server.py"],      "env": {        "LLM_API_KEY": "your-api-key",        "LLM_BASE_URL": "https://api.openai.com/v1",        "LLM_MODEL": "gpt-4o-mini"      }    }  }}

Cursor / Windsurf

在 IDE 的 MCP 配置中添加上述 stdio 配置即可。

HTTP/SSE 客户端

配置 MCP 服务器地址为 http://localhost:8000/mcp。

📁 目录结构

contract-risk-review-mcp/├── server.py              # MCP Server 主文件├── requirements.txt       # 依赖清单├── README.md              # 使用说明└── knowledge/             # 知识库    ├── risk-knowledge-base.md    # 劳动关系16类风险知识库    ├── service-contract-risks.md # 劳务关系9类风险知识库    ├── local-regulations.md      # 地方法规差异知识库    └── categories-index.md       # 风险类别索引

⚠️ 免责声明

本工具仅提供合同风险参考、条款修改建议以及法规依据，不构成正式法律意见，不能替代执业律师的专业审核。涉及劳动仲裁、重大权益纠纷，请务必咨询执业律师。

📝 更新日志

v1.0.0

  • 初始版本
  • 支持劳动合同 16 类风险审查
  • 支持劳务合同 9 类风险审查
  • 支持批量审查、知识库查询、地方法规查询
  • 支持 STDIO 和 HTTP/SSE 两种传输模式

Source: README.md at commit aaa88da

Tools

0
SourceWeft has not started this package in its sandbox yet, so its tools are not known.

Version history

1
  1. v1.0.0LatestSep 16, 2026

Score

40% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts

    Declare a license in the repository or the registry entry.

Other Review integrations

Review Expert

expert.reviewOfficial publisher

Lets an assistant read, search, analyze and draft replies to App Store and Google Play reviews, with OAuth sign-in.

Streamable HTTPWeb executableAuth
Communication & CollaborationCredential requiredOct 24, 2026

DECIDE: Migration Discovery Risk Review

Published by GitHub account stratus-consulting

A remote risk-review service that takes a migration's discovery findings and returns open risks, highest severity first, each with who to ask next.

Streamable HTTPWeb executable
Productivity & Workflow1 toolNov 7, 2026

Arch theological review

Published by GitHub account tnvolman

Cursor/Grok plugin giving an assistant a paid Arch theological review of faith text and a free offline King James verse lookup.

STDIODesktop only
Knowledge & MemoryNot validated · the package type cannot run in the sandboxNov 6, 2026

ReviewBot App Reviews

Published by GitHub account fahimk

Read collected App Store and Google Play reviews and bounded feedback statistics.

Streamable HTTPWeb executableAuth
Credential requiredNov 4, 2026

Khavion Architecture Review

com.khavion

Local MCP server that gives an assistant cited cloud best-practice quotes and a deterministic rules-based review of Terraform, CloudFormation and Kubernetes…

STDIODesktop only
Cloud & Infrastructure4 toolsNov 4, 2026