AI合规检查服务器

Gibs MCP Server是一款AI合规检查服务器,用于分类AI系统、检查合规义务并提供法规级别的引用,覆盖240+条款和3个主要法规。

  • by gibbrdev
  • 云端部署
  • 7 星标
  • 更新于
在 GitHub 查看 buildsyncinc/gibs-mcp

使用说明

Gibs MCP服务器

人工智能开发工具的监管合规性 --对人工智能系统进行分类,检查义务,并在240多篇文章和3项主要法规中获得文章级引用。

全部文件 · 获取API密钥 · API 参考


概述

Gibs MCP Server将您的AI开发环境直接连接到法规遵从性知识库。目前涵盖:

法规 范围 条款
欧盟人工智能法案 人工智能系统分类、禁止行为、风险等级义务 113条+13个附件
GDPR 数据保护 数据保护、处理义务、数据主体权利 99篇
多拉 ICT风险管理、事件报告、金融实体第三方监督 64篇文章+12项授权/实施法案

每个回复都包括 文章级引用 随着法规的发展,实时更新语料库,以约束法律文本。


快速开始

1.获取API密钥

注册地址: gibs.dev 从仪表板中获取API密钥。

2.通过克劳德桌面连接

将此添加到您的Claude Desktop配置文件中(claude_desktop_config.json):

{
  "mcpServers": {
    "gibs": {
      "url": "https://mcp.gibs.dev/sse"
    }
  }
}

3.身份验证

将您的API密钥作为 user_api_key 调用任何工具时的参数,或在您的环境中设置它:

export GIBS_API_KEY=your_api_key_here

工具

classify_ai_system

根据《人工智能法案》的风险等级(不可接受、高风险、有限、最低)对人工智能系统进行分类,并进行充分的法律推理。

参数 类型 必填 说明
description string 是 人工智能系统做什么(10-5000个字符)
data_types list[string] 否 处理的数据类型(例如。, ["biometric", "personal"])
decision_scope string 否 系统会影响哪些决策
sector string 否 行业部门(例如。, "healthcare", "finance", "hr")

示例提示:

为学校分类我的面部识别考勤系统

你会得到什么:

  • 风险分类(例如。, 高危的 附件三第1(a)点)
  • 具体的第六条分析,并附有附件三的交叉引用
  • 基于分类的适用义务
  • 相关豁免或例外情况(如有)

check_compliance

询问任何监管合规问题,并获得文章级引用的结构化答案。

参数 类型 必填 说明
question string 是 合规性问题(10-2000个字符)
system_description string 否 关于您的AI系统或组织的上下文
regulation string 否 目标规则: "ai_act", "gdpr", "dora",或 "both" (默认值: "both")

示例提示:

GDPR对数据泄露通知的要求是什么?

返回第33条(通知监管机构)和第34条(与数据主体的沟通)的义务,包括72小时的通知时间、所需内容和例外情况。

DORA是否适用于处理付款的金融科技初创公司?

返回第2条范围分析,包括金融实体分类、适用的ICT风险管理义务和相关委托行为要求。


health_check

检查API合规性是否可用。无需参数。


使用示例

1.人工智能系统分类

You: Classify my facial recognition attendance system for schools

Gibs: This system is classified as HIGH-RISK under the EU AI Act.

- Article 6(2) read with Annex III, point 1(a): biometric identification
  and categorisation of natural persons
- As used in an educational context, also caught by Annex III, point 3(a):
  AI systems intended to be used to determine access to educational institutions

Obligations include:
- Risk management system (Article 9)
- Data governance (Article 10)
- Technical documentation (Article 11)
- Record-keeping (Article 12)
- Transparency to users (Article 13)
- Human oversight (Article 14)
- Accuracy, robustness, cybersecurity (Article 15)
- Conformity assessment before placing on market (Article 43)

2.GDPR数据泄露义务

You: What are the GDPR requirements for a data breach notification?

Gibs: GDPR imposes two notification obligations following a personal data breach:

1. Notification to supervisory authority — Article 33
   - Without undue delay and no later than 72 hours after becoming aware
   - Must include: nature of breach, categories and approximate number of
     data subjects, likely consequences, measures taken or proposed
   - Exception: unlikely to result in a risk to rights and freedoms

2. Communication to data subjects — Article 34
   - Required when breach is likely to result in a HIGH risk
   - Must describe in clear and plain language the nature of the breach
   - Not required if: appropriate technical safeguards applied (e.g.,
     encryption), subsequent measures taken that eliminate the high risk,
     or it would involve disproportionate effort (use public communication instead)

The controller must also document all breaches per Article 33(5), regardless
of whether they meet the notification threshold.

3.DORA范围分析

You: Does DORA apply to a fintech startup processing payments?

Gibs: Yes. Under Article 2(1), DORA applies to payment institutions as
defined in Directive 2015/2366 (PSD2), which includes entities authorized
to provide payment services.

As a financial entity under DORA, the startup must comply with:
- ICT risk management framework (Chapter II, Articles 5–16)
- ICT-related incident management and reporting (Chapter III, Articles 17–23)
- Digital operational resilience testing (Chapter IV, Articles 24–27)
- ICT third-party risk management (Chapter V, Articles 28–44)
- Information-sharing arrangements (Chapter VI, Article 45)

Note: Proportionality applies — Article 4 allows competent authorities to
consider the size, risk profile, and complexity of the financial entity when
assessing compliance.

连接详情

SSE端点 https://mcp.gibs.dev/sse
协议 模型上下文协议(MCP) 服务器发送的事件
认证 API密钥来自 gibs.dev 仪表板

支持的客户

任何兼容MCP的客户端都可以连接到Gibs服务器。测试方法:


链接


许可证

MIT许可证。看 许可证 了解详情。


建造于 Gibbr AB 公司 --使开发人员能够遵守监管规定。