ISO 42001 AI MCP
ISO 42001 AI MCP
针对 ISO/IEC 42001:2023 的人工智能管理系统 (AIMS) 评估、认证准备及欧盟《人工智能法案》(EU AI Act) 对照分析。
为什么存在此项目
ISO 42001:2023 是全球首个专门针对人工智能管理系统的认证标准。它定义了组织在负责任地开发、部署和维护 AI 时必须采取的措施。认证机构(如 BSI、TUV、SGS、必维国际检验集团)目前正在颁发 ISO 42001 证书,企业在采购过程中也开始要求此认证。
该标准包含 39 项附录 A 控制措施和 9 项管理体系条款。将这些内容映射到您的 AI 生命周期、与欧盟《人工智能法案》的合规性评估进行对照,并为第一阶段/第二阶段审核做准备,通常需要 2-5 万美元的咨询费用。此 MCP 可通过单个提示词执行完整的 AIMS 评估、风险分析、策略生成、附录 A 控制检查、欧盟《人工智能法案》对照分析以及认证时间表规划。
Related MCP server: ISO 27001 AI MCP
安装
pip install iso-42001-ai-mcp工具
工具 | ISO 参考 | 功能描述 |
| 第 4-10 条款 | 针对 ISO 42001:2023 管理条款的完整 AIMS 审核 |
| 第 6.1 条款 | 具有影响力和可能性评分的 AI 专项风险评估 |
| 第 5.2 条款,附录 A | 生成符合管理承诺要求的 AI 策略 |
| 附录 A (39 项控制) | 对所有附录 A 目标的逐项控制评估 |
| 附录 A + 欧盟《人工智能法案》 | 将 ISO 42001 控制措施映射到欧盟《人工智能法案》的合规要求 |
| 第一阶段 / 第二阶段 | 包含时间表的认证准备清单 |
| 机器学习辅助 | 用于 AI 系统的神经网络风险预测 |
| 所有条款 | 快速 AI 系统合规性概览 |
| 完整生命周期 | 第一阶段/第二阶段审核时间表及里程碑 |
示例
Prompt: "Assess our computer vision system for ISO 42001 certification
readiness. It processes facial images for building access control,
was trained on a proprietary dataset, and has no explainability layer."
Result: AIMS assessment with findings across Annex A controls: biometric
processing triggers A.6.2.4 (impact assessment), missing explainability
fails A.6.2.6 (transparency), proprietary dataset needs A.7.3 (data
management). EU AI Act crosswalk flags Annex III high-risk classification.
Certification timeline generated with 14-week remediation path.定价
层级 | 价格 | 您将获得 |
免费版 | £0 | 每天 10 次调用 — AIMS 审核 + 快速扫描 |
专业版 | £199/月 | 无限调用 + HMAC 签名认证 + 验证 URL |
企业版 | £1,499/月 | 多租户 + 联合品牌报告 + Webhooks |
认证 API
每次专业版/企业版审核都会生成一份经过加密签名的证书:
POST https://meok-attestation-api.vercel.app/sign
GET https://meok-attestation-api.vercel.app/verify/{cert_id}零依赖验证器:pip install meok-attestation-verify
链接
网站:meok.ai
所有 MCP 服务器:meok.ai/labs/mcp/servers
企业支持:nicholas@csoai.org
许可证
MIT
Available Tools
10 toolsassess_ai_riskA
Perform ISO 42001 Annex B risk assessment for AI systems.
Comprehensive AI risk assessment covering risk criteria establishment, risk identification across all Annex B categories, risk analysis (likelihood and impact), and risk evaluation against organizational risk criteria. Follows the ISO 42001 Annex B guidance structure.
Args: system_description: Detailed description of the AI system including purpose, data, deployment context, and affected populations. system_name: Name of the AI system. risk_criteria: Organization's risk acceptance criteria description. caller: Caller identifier for rate limiting. tier: Pricing tier ('free' or 'pro').
Returns: Complete Annex B risk assessment with identified risks, analysis, and evaluation results.
Behavior: This tool is read-only and stateless — it produces analysis output without modifying any external systems, databases, or files. Safe to call repeatedly with identical inputs (idempotent). Free tier: 10/day rate limit. Pro tier: unlimited. No authentication required for basic usage.
When to use: Use this tool when you need to assess, audit, or verify compliance requirements. Ideal for gap analysis, readiness checks, and generating compliance documentation.
When NOT to use: Do not use as a substitute for qualified legal counsel. This tool provides technical compliance guidance, not legal advice. Behavioral Transparency: - Side Effects: This tool is read-only and produces no side effects. It does not modify any external state, databases, or files. All output is computed in-memory and returned directly to the caller. - Authentication: No authentication required for basic usage. Pro/Enterprise tiers require a valid MEOK API key passed via the MEOK_API_KEY environment variable. - Rate Limits: Free tier: 10 calls/day. Pro tier: unlimited. Rate limit headers are included in responses (X-RateLimit-Remaining, X-RateLimit-Reset). - Error Handling: Returns structured error objects with 'error' key on failure. Never raises unhandled exceptions. Invalid inputs return descriptive validation errors. - Idempotency: Fully idempotent — calling with the same inputs always produces the same output. Safe to retry on timeout or transient failure. - Data Privacy: No input data is stored, logged, or transmitted to external services. All processing happens locally within the MCP server process.
Inputs: system_description (required), system_name, risk_criteria, caller, +2 more.
audit_management_systemA
Audit an AI management system against ISO/IEC 42001 clauses 4-10.
Evaluates organizational readiness and conformity across all seven management system clauses: Context (4), Leadership (5), Planning (6), Support (7), Operation (8), Performance Evaluation (9), and Improvement (10). Returns per-clause assessment with audit questions, gap analysis, and prioritized recommendations.
Args: organization_description: Description of the organization and its AI management practices, governance structures, and policies. ai_systems_description: Description of AI systems in scope. existing_certifications: Existing ISO or other certifications held (e.g., 'ISO 27001, ISO 9001'). caller: Caller identifier for rate limiting. tier: Pricing tier ('free' or 'pro').
Returns: Clause-by-clause audit results with conformity status and recommendations.
Behavior: This tool is read-only and stateless — it produces analysis output without modifying any external systems, databases, or files. Safe to call repeatedly with identical inputs (idempotent). Free tier: 10/day rate limit. Pro tier: unlimited. No authentication required for basic usage.
When to use: Use this tool when you need to assess, audit, or verify compliance requirements. Ideal for gap analysis, readiness checks, and generating compliance documentation.
When NOT to use: Do not use as a substitute for qualified legal counsel. This tool provides technical compliance guidance, not legal advice. Behavioral Transparency: - Side Effects: This tool is read-only and produces no side effects. It does not modify any external state, databases, or files. All output is computed in-memory and returned directly to the caller. - Authentication: No authentication required for basic usage. Pro/Enterprise tiers require a valid MEOK API key passed via the MEOK_API_KEY environment variable. - Rate Limits: Free tier: 10 calls/day. Pro tier: unlimited. Rate limit headers are included in responses (X-RateLimit-Remaining, X-RateLimit-Reset). - Error Handling: Returns structured error objects with 'error' key on failure. Never raises unhandled exceptions. Invalid inputs return descriptive validation errors. - Idempotency: Fully idempotent — calling with the same inputs always produces the same output. Safe to retry on timeout or transient failure. - Data Privacy: No input data is stored, logged, or transmitted to external services. All processing happens locally within the MCP server process.
Inputs: organization_description (required), ai_systems_description, existing_certifications, caller, +2 more.
certification_timelineA
Returns ISO 42001 certification steps and typical timelines. No parameters needed.
Behavior: This tool is read-only and stateless — it produces analysis output without modifying any external systems, databases, or files. Safe to call repeatedly with identical inputs (idempotent). Free tier: 10/day rate limit. Pro tier: unlimited. No authentication required for basic usage.
When to use: Use this tool when you need to assess, audit, or verify compliance requirements. Ideal for gap analysis, readiness checks, and generating compliance documentation.
When NOT to use: Do not use as a substitute for qualified legal counsel. This tool provides technical compliance guidance, not legal advice. Behavioral Transparency: - Side Effects: This tool is read-only and produces no side effects. It does not modify any external state, databases, or files. All output is computed in-memory and returned directly to the caller. - Authentication: No authentication required for basic usage. Pro/Enterprise tiers require a valid MEOK API key passed via the MEOK_API_KEY environment variable. - Rate Limits: Free tier: 10 calls/day. Pro tier: unlimited. Rate limit headers are included in responses (X-RateLimit-Remaining, X-RateLimit-Reset). - Error Handling: Returns structured error objects with 'error' key on failure. Never raises unhandled exceptions. Invalid inputs return descriptive validation errors. - Idempotency: Fully idempotent — calling with the same inputs always produces the same output. Safe to retry on timeout or transient failure. - Data Privacy: No input data is stored, logged, or transmitted to external services. All processing happens locally within the MCP server process.
Inputs: no named parameters.
check_annex_controlsA
Evaluate AI system against ISO 42001 Annex A controls.
Maps the system to all Annex A control objectives and evaluates which controls are applicable and their implementation status. Produces a gap analysis suitable for Statement of Applicability.
Args: system_description: Description of the AI system and its management. system_name: Name of the AI system. implemented_controls: Description of controls already implemented (free text or comma-separated control IDs). caller: Caller identifier for rate limiting. tier: Pricing tier ('free' or 'pro').
Returns: Annex A control evaluation with applicability and gap analysis.
Behavior: This tool generates structured output without modifying external systems. Output is deterministic for identical inputs. No side effects. Free tier: 10/day rate limit. Pro tier: unlimited. No authentication required for basic usage.
When to use: Use this tool when you need to assess, audit, or verify compliance requirements. Ideal for gap analysis, readiness checks, and generating compliance documentation.
When NOT to use: Do not use as a substitute for qualified legal counsel. This tool provides technical compliance guidance, not legal advice. Behavioral Transparency: - Side Effects: This tool is read-only and produces no side effects. It does not modify any external state, databases, or files. All output is computed in-memory and returned directly to the caller. - Authentication: No authentication required for basic usage. Pro/Enterprise tiers require a valid MEOK API key passed via the MEOK_API_KEY environment variable. - Rate Limits: Free tier: 10 calls/day. Pro tier: unlimited. Rate limit headers are included in responses (X-RateLimit-Remaining, X-RateLimit-Reset). - Error Handling: Returns structured error objects with 'error' key on failure. Never raises unhandled exceptions. Invalid inputs return descriptive validation errors. - Idempotency: Fully idempotent — calling with the same inputs always produces the same output. Safe to retry on timeout or transient failure. - Data Privacy: No input data is stored, logged, or transmitted to external services. All processing happens locally within the MCP server process.
Inputs: system_description (required), system_name, implemented_controls, caller, +2 more.
create_certification_checklistA
Generate ISO 42001 certification readiness checklist with pass/fail.
Creates a comprehensive certification preparation checklist covering all mandatory requirements, documentation, and evidence needed for ISO 42001 certification audit. Includes pre-audit assessment and remediation guidance.
Args: organization_name: Name of the organization pursuing certification. current_status: Description of current AIMS implementation status. target_date: Target certification date (YYYY-MM-DD format). caller: Caller identifier for rate limiting. tier: Pricing tier ('free' or 'pro').
Returns: Certification readiness checklist with pass/fail status per item.
Behavior: This tool generates structured output without modifying external systems. Output is deterministic for identical inputs. No side effects. Free tier: 10/day rate limit. Pro tier: unlimited. No authentication required for basic usage.
When to use: Use this tool when you need to assess, audit, or verify compliance requirements. Ideal for gap analysis, readiness checks, and generating compliance documentation.
When NOT to use: Do not use as a substitute for qualified legal counsel. This tool provides technical compliance guidance, not legal advice. Behavioral Transparency: - Side Effects: This tool is read-only and produces no side effects. It does not modify any external state, databases, or files. All output is computed in-memory and returned directly to the caller. - Authentication: No authentication required for basic usage. Pro/Enterprise tiers require a valid MEOK API key passed via the MEOK_API_KEY environment variable. - Rate Limits: Free tier: 10 calls/day. Pro tier: unlimited. Rate limit headers are included in responses (X-RateLimit-Remaining, X-RateLimit-Reset). - Error Handling: Returns structured error objects with 'error' key on failure. Never raises unhandled exceptions. Invalid inputs return descriptive validation errors. - Idempotency: Fully idempotent — calling with the same inputs always produces the same output. Safe to retry on timeout or transient failure. - Data Privacy: No input data is stored, logged, or transmitted to external services. All processing happens locally within the MCP server process.
Inputs: organization_name, current_status, target_date, caller, +2 more.
crosswalk_to_eu_ai_actA
Map ISO/IEC 42001 clauses and Annex A controls to EU AI Act articles.
This is the killer feature -- regulation-to-regulation mapping showing exactly where ISO 42001 conformity satisfies EU AI Act requirements. Essential for organizations pursuing ISO 42001 certification while preparing for EU AI Act compliance.
Args: iso_clauses: Comma-separated ISO clauses to crosswalk (e.g., '4,5,8') or 'all' for complete mapping. Include 'annex' for Annex A mappings. focus_area: Optional focus area to filter (e.g., 'risk management', 'transparency', 'data governance', 'human oversight'). caller: Caller identifier for rate limiting. tier: Pricing tier ('free' or 'pro').
Returns: Detailed crosswalk between ISO 42001 and EU AI Act with alignment strength ratings and dual-compliance guidance.
Behavior: This tool is read-only and stateless — it produces analysis output without modifying any external systems, databases, or files. Safe to call repeatedly with identical inputs (idempotent). Free tier: 10/day rate limit. Pro tier: unlimited. No authentication required for basic usage.
When to use: Use this tool when you need to assess, audit, or verify compliance requirements. Ideal for gap analysis, readiness checks, and generating compliance documentation.
When NOT to use: Do not use as a substitute for qualified legal counsel. This tool provides technical compliance guidance, not legal advice. Behavioral Transparency: - Side Effects: This tool is read-only and produces no side effects. It does not modify any external state, databases, or files. All output is computed in-memory and returned directly to the caller. - Authentication: No authentication required for basic usage. Pro/Enterprise tiers require a valid MEOK API key passed via the MEOK_API_KEY environment variable. - Rate Limits: Free tier: 10 calls/day. Pro tier: unlimited. Rate limit headers are included in responses (X-RateLimit-Remaining, X-RateLimit-Reset). - Error Handling: Returns structured error objects with 'error' key on failure. Never raises unhandled exceptions. Invalid inputs return descriptive validation errors. - Idempotency: Fully idempotent — calling with the same inputs always produces the same output. Safe to retry on timeout or transient failure. - Data Privacy: No input data is stored, logged, or transmitted to external services. All processing happens locally within the MCP server process.
Inputs: iso_clauses, focus_area, caller, tier, +1 more.
generate_policy_templateA
Generate AI policy documents per ISO 42001 requirements.
Creates policy templates that satisfy ISO 42001 clause 5.2 (AI policy) and Annex A.2.2 requirements. Includes AI policy statement, roles and responsibilities, objectives, principles, and governance structure.
Args: organization_name: Name of the organization. ai_scope: Description of AI systems and activities in scope. policy_type: Type of policy ('comprehensive', 'brief', 'executive'). caller: Caller identifier for rate limiting. tier: Pricing tier ('free' or 'pro').
Returns: Markdown-formatted policy template with all required elements.
Behavior: This tool generates structured output without modifying external systems. Output is deterministic for identical inputs. No side effects. Free tier: 10/day rate limit. Pro tier: unlimited. No authentication required for basic usage.
When to use: Use this tool when you need to assess, audit, or verify compliance requirements. Ideal for gap analysis, readiness checks, and generating compliance documentation.
When NOT to use: Do not use as a substitute for qualified legal counsel. This tool provides technical compliance guidance, not legal advice. Behavioral Transparency: - Side Effects: This tool is read-only and produces no side effects. It does not modify any external state, databases, or files. All output is computed in-memory and returned directly to the caller. - Authentication: No authentication required for basic usage. Pro/Enterprise tiers require a valid MEOK API key passed via the MEOK_API_KEY environment variable. - Rate Limits: Free tier: 10 calls/day. Pro tier: unlimited. Rate limit headers are included in responses (X-RateLimit-Remaining, X-RateLimit-Reset). - Error Handling: Returns structured error objects with 'error' key on failure. Never raises unhandled exceptions. Invalid inputs return descriptive validation errors. - Idempotency: Fully idempotent — calling with the same inputs always produces the same output. Safe to retry on timeout or transient failure. - Data Privacy: No input data is stored, logged, or transmitted to external services. All processing happens locally within the MCP server process.
Inputs: organization_name (required), ai_scope, policy_type, caller, +2 more.
neural_insightsA
Get aggregate learning insights from the neural compliance model.
Behavior: This tool is read-only and stateless — it produces analysis output without modifying any external systems, databases, or files. Safe to call repeatedly with identical inputs (idempotent). Free tier: 10/day rate limit. Pro tier: unlimited. No authentication required for basic usage.
When to use: Use this tool when you need to assess, audit, or verify compliance requirements. Ideal for gap analysis, readiness checks, and generating compliance documentation.
When NOT to use: Do not use as a substitute for qualified legal counsel. This tool provides technical compliance guidance, not legal advice. Behavioral Transparency: - Side Effects: This tool is read-only and produces no side effects. It does not modify any external state, databases, or files. All output is computed in-memory and returned directly to the caller. - Authentication: No authentication required for basic usage. Pro/Enterprise tiers require a valid MEOK API key passed via the MEOK_API_KEY environment variable. - Rate Limits: Free tier: 10 calls/day. Pro tier: unlimited. Rate limit headers are included in responses (X-RateLimit-Remaining, X-RateLimit-Reset). - Error Handling: Returns structured error objects with 'error' key on failure. Never raises unhandled exceptions. Invalid inputs return descriptive validation errors. - Idempotency: Fully idempotent — calling with the same inputs always produces the same output. Safe to retry on timeout or transient failure. - Data Privacy: No input data is stored, logged, or transmitted to external services. All processing happens locally within the MCP server process.
Inputs: api_key.
predict_risk_neuralA
Neural network-based risk prediction that improves from every compliance check.
Behavior: This tool is read-only and stateless — it produces analysis output without modifying any external systems, databases, or files. Safe to call repeatedly with identical inputs (idempotent). Free tier: 10/day rate limit. Pro tier: unlimited. No authentication required for basic usage.
When to use: Use this tool when you need to assess, audit, or verify compliance requirements. Ideal for gap analysis, readiness checks, and generating compliance documentation.
When NOT to use: Do not use as a substitute for qualified legal counsel. This tool provides technical compliance guidance, not legal advice. Behavioral Transparency: - Side Effects: This tool is read-only and produces no side effects. It does not modify any external state, databases, or files. All output is computed in-memory and returned directly to the caller. - Authentication: No authentication required for basic usage. Pro/Enterprise tiers require a valid MEOK API key passed via the MEOK_API_KEY environment variable. - Rate Limits: Free tier: 10 calls/day. Pro tier: unlimited. Rate limit headers are included in responses (X-RateLimit-Remaining, X-RateLimit-Reset). - Error Handling: Returns structured error objects with 'error' key on failure. Never raises unhandled exceptions. Invalid inputs return descriptive validation errors. - Idempotency: Fully idempotent — calling with the same inputs always produces the same output. Safe to retry on timeout or transient failure. - Data Privacy: No input data is stored, logged, or transmitted to external services. All processing happens locally within the MCP server process.
Inputs: system_name (required), uses_biometric, uses_health_data, has_human_oversight, +5 more.
quick_scanA
One-line system description to instant ISO 42001 gap assessment. No API key needed.
Args: description: Brief description of your AI system (e.g. 'customer service chatbot using GPT-4')
Returns: Instant gap assessment against ISO/IEC 42001:2023 clauses.
Behavior: This tool is read-only and stateless — it produces analysis output without modifying any external systems, databases, or files. Safe to call repeatedly with identical inputs (idempotent). Free tier: 10/day rate limit. Pro tier: unlimited. No authentication required for basic usage.
When to use: Use this tool when you need to assess, audit, or verify compliance requirements. Ideal for gap analysis, readiness checks, and generating compliance documentation.
When NOT to use: Do not use as a substitute for qualified legal counsel. This tool provides technical compliance guidance, not legal advice. Behavioral Transparency: - Side Effects: This tool is read-only and produces no side effects. It does not modify any external state, databases, or files. All output is computed in-memory and returned directly to the caller. - Authentication: No authentication required for basic usage. Pro/Enterprise tiers require a valid MEOK API key passed via the MEOK_API_KEY environment variable. - Rate Limits: Free tier: 10 calls/day. Pro tier: unlimited. Rate limit headers are included in responses (X-RateLimit-Remaining, X-RateLimit-Reset). - Error Handling: Returns structured error objects with 'error' key on failure. Never raises unhandled exceptions. Invalid inputs return descriptive validation errors. - Idempotency: Fully idempotent — calling with the same inputs always produces the same output. Safe to retry on timeout or transient failure. - Data Privacy: No input data is stored, logged, or transmitted to external services. All processing happens locally within the MCP server process.
Inputs: description (required).
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 10 tools
Most tools have distinct purposes, but there is overlap between risk assessment tools (assess_ai_risk, predict_risk_neural) and between audit/checklist tools (audit_management_system, create_certification_checklist). The vague descriptions of neural_insights and predict_risk_neural add confusion.
Most tools use verb_noun snake_case, but neural_insights (noun_noun) and quick_scan (adjective_noun) deviate. predict_risk_neural also inverts the expected order. The pattern is not fully consistent.
10 tools cover the core domain of ISO 42001 compliance well. The count is reasonable, though a few tools (e.g., neural_insights, predict_risk_neural) could be merged or clarified without loss.
The tool set covers risk assessment, management system audit, certification, Annex A controls, policy generation, and crosswalk to EU AI Act. Minor gaps exist, such as no direct support for corrective actions or evidence management.
Maintenance
Related MCP Connectors
AI governance MCP server for EU AI Act compliance and jurisdiction verification
ISO/IEC 42005:2025 AI Impact Assessment MCP — first international standard for AI impact
Agent Orchestrator MCP Server by MEOK AI Labs
A2a Governance Bridge MCP Server by MEOK AI Labs
Related MCP Servers
- AlicenseNot gradedqualityCmaintenanceNIST RMF AI - MCP server providing AI-powered tools and automation by MEOK AI Labs2 npm42 PyPIMIT

ISO 27001 AI MCPofficial
AlicenseNot gradedqualityCmaintenanceISO 27001 AI - MCP server providing AI-powered tools and automation by MEOK AI Labs3 npm43 PyPI3MIT- AlicenseAqualityCmaintenanceEU AI Act Compliance - MCP server providing AI-powered tools and automation by MEOK AI Labs111MIT

SOC2 Compliance AI MCPofficial
AlicenseAqualityCmaintenanceSOC2 Compliance AI - MCP server providing AI-powered tools and automation by MEOK AI Labs66 npm70 PyPIMIT