智能SFC控制平面
智能SFC控制平面 是 AI Skill Hub 本期精选MCP工具之一。综合评分 7.5 分，整体质量较高。我们推荐使用将其纳入你的 AI 工具库，帮助提升工作效率。
📚 深度解析
通过安装 智能SFC控制平面，你的 AI 助手将获得额外的工具调用能力，可以用自然语言直接操控该工具的功能，无需学习复杂的命令行语法。MCP 工具的核心价值在于"一次配置，永久增强"——配置完成后，每次与 AI 对话时都可以无缝调用这些工具。
在技术实现上，MCP 工具通过标准的 JSON-RPC 协议与 AI 客户端通信，工具的功能以"工具列表"的形式暴露给 AI 模型，AI 可以按需调用。智能SFC控制平面 提供了结构化的工具调用接口，使 AI 模型能够精确地理解和使用每个功能点，显著降低 AI 在工具使用上的错误率。
与传统的 API 集成相比，MCP 工具的优势在于无需编写代码——用户只需在配置文件中添加几行 JSON，即可让 AI 获得全新能力。AI Skill Hub 将 智能SFC控制平面 评为 AI 评分 7.5 分，属于同类工具中的优质选择。
📋 工具概览
智能SFC控制平面 是一款遵循 MCP（Model Context Protocol）标准协议的 AI 工具扩展。通过 MCP 协议，它可以让 Claude、Cursor 等主流 AI 客户端直接访问和操作外部工具、数据源和服务，实现 AI 能力的无缝扩展。无论是文件操作、数据库查询还是 API 调用，都可以通过自然语言在 AI 对话中直接触发，极大提升生产效率。
📖 中文文档
智能SFC控制平面 是一款遵循 MCP（Model Context Protocol）标准协议的 AI 工具扩展。通过 MCP 协议，它可以让 Claude、Cursor 等主流 AI 客户端直接访问和操作外部工具、数据源和服务，实现 AI 能力的无缝扩展。无论是文件操作、数据库查询还是 API 调用，都可以通过自然语言在 AI 对话中直接触发，极大提升生产效率。
- 通过标准 MCP 协议与 Claude、Cursor 等主流 AI 客户端深度集成
- 提供结构化工具调用接口，显著降低 AI 集成复杂度
- 支持 Claude Desktop 和 Claude Code 无缝接入，开箱即用
- 可与其他 MCP 工具组合叠加，构建完整 AI 工作站
- 轻量无侵入设计，不影响现有系统架构
- 在 Claude Desktop 对话中直接调用本地工具，实现 AI 与系统的深度联动
- 通过自然语言驱动复杂的多步骤自动化任务，代替繁琐手动操作
- 将多个 MCP 工具组合使用，构建个人专属 AI 工作站
# 方式一：通过 Claude Code CLI 一键安装
claude skill install https://github.com/aws-samples/sample-sfc-agentic-control-plane
# 方式二：手动配置 claude_desktop_config.json
{
"mcpServers": {
"--sfc----": {
"command": "npx",
"args": ["-y", "sample-sfc-agentic-control-plane"]
}
}
}
# 配置文件位置
# macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
# Windows: %APPDATA%/Claude/claude_desktop_config.json
- 确认已安装 Node.js（v18 或以上版本）
- 打开 Claude Desktop 或 Claude Code 的 MCP 配置文件
- 按「交给 Agent 安装 → Claude Desktop」标签中的 JSON 配置填入 mcpServers 字段
- 保存配置文件并重启 Claude 客户端
- 重启后，在对话中即可使用本工具
# 安装后在 Claude 对话中直接使用 # 示例： 用户: 请帮我用 智能SFC控制平面 执行以下任务... Claude: [自动调用 智能SFC控制平面 MCP 工具处理请求] # 查看可用工具列表 # 在 Claude 中输入："列出所有可用的 MCP 工具"
// claude_desktop_config.json 配置示例
{
"mcpServers": {
"__sfc____": {
"command": "npx",
"args": ["-y", "sample-sfc-agentic-control-plane"],
"env": {
// "API_KEY": "your-api-key-here"
}
}
}
}
// 保存后重启 Claude Desktop 生效
SFC Agentic Control Plane
<img src="docs/sfc-control-plane-logo.svg" alt="SFC Control Plane Logo" width="64" height="64" />
---
Text descriptions
Executive Summary
Connecting industrial equipment to cloud data pipelines is one of manufacturing's most persistent bottlenecks. The SFC Agentic Control Plane eliminates this barrier by combining a conversational AI assistant with a production-grade cloud control plane. Engineers describe what they need — in plain language or by uploading existing machine specs — and the agent produces a validated, deployment-ready Shop Floor Connectivity (SFC) configuration. That configuration is then packaged, cryptographically credentialed, and pushed to the edge in a single click. If the running process emits errors, a second AI step diagnoses the logs and proposes a corrected configuration automatically.
#### Pitch The SFC Agentic Control Plane eliminates the barrier of onboarding industrial equipment by combining an LLM Agent with a production-grade cloud control plane.
---
Abstract
This solution wraps AWS Shop Floor Connectivity (SFC) — with an AI-driven lifecycle. The SFC Config Agent runs as an Amazon Bedrock AgentCore Runtime backed by Claude on Amazon Bedrock. It uses a purpose-built MCP server to validate configurations against the live SFC specification before saving them. A serverless SFC Control Plane (API Gateway + Lambda + DynamoDB + S3) stores versioned configs, assembles self-contained "Launch Packages" complete with AWS IoT X.509 credentials, and streams OpenTelemetry logs from the edge back to CloudWatch. A React/TypeScript single-page app (SPA) served via CloudFront ties all of this together into an operator-facing workflow that goes from an empty text box to a monitored, remotely-controllable edge process in minutes.
Capabilities & Ideas
The core idea is that SFC configuration is expert knowledge that most OT engineers lack and most IT teams don't have time to acquire. By grounding an LLM in the actual SFC specification — via an MCP server that reads directly from the SFC GitHub repository — the agent generates correct-by-construction configs rather than plausible-looking but broken JSON. Every generated config is validated by the same MCP tools before it is persisted, creating a tight correctness loop that does not rely on model memorization.
The control plane extends this idea to the full device lifecycle. A "Config in Focus" concept — a pinned config version displayed prominently in the UI — makes it unambiguous which configuration will be used the next time a Launch Package is created. Launch Packages are self-contained zip archives that embed the SFC config, an AWS IoT-provisioned X.509 device certificate, a role alias for temporary AWS credential vending, and a runtime agent (aws-sfc-runtime-agent). Operators download the zip, unpack it on any Windows, Mac or Linux host, and run a single command. No cloud credentials are baked in; the edge device exchanges its device certificate for short-lived IAM credentials on every session via the AWS IoT Credential Provider.
Remote operations are handled over an MQTT5 control channel. From the UI, operators can toggle OpenTelemetry log shipping on or off, switch SFC to TRACE-level diagnostics, push a new config version over-the-air, or trigger a graceful SFC restart — all without touching the edge host. A live heartbeat LED (green / red / grey) reflects device status at a glance, updated every ten seconds.
AI-assisted remediation closes the loop. When ERROR-severity OTEL records appear in the log viewer, a single "Fix with AI" click sends the error window and the current SFC config to the agent. The agent diagnoses the errors using its MCP-backed SFC knowledge, returns a corrected config, and the UI renders a side-by-side diff. One more click creates a new Launch Package from the corrected config, preserving the full lineage chain back to the failed deployment.
The result is an end-to-end workflow — from natural-language description to monitored, self-healing edge process — built entirely on AWS serverless primitives with no standing infrastructure costs.
---
`ui/.env.local` — required variables
After cdk deploy, copy the CDK outputs into src/ui/.env.local:
```dotenv
Deployment & Quickstart
1. Deploy the CDK Stack
npm install
npx cdk deploy -c region=<YOUR_REGION>Default region: If-c region=is omitted, the stack falls back to theCDK_DEFAULT_REGIONenvironment variable, then tous-east-1.
The CDK stack provisions all infrastructure and: 1. Uploads local sources to S3 for CodeBuild 2. Triggers the AgentCore deployment CodeBuild project (builds and registers the AI agent container) 3. Triggers the UI build CodeBuild project (runs npm run build for the Vite SPA and syncs assets to S3) 4. Serves the UI via Amazon CloudFront — the URL is printed as SfcControlPlaneUiUrl
Key CDK outputs:
| Output | Description |
|---|---|
SfcControlPlaneUiUrl | CloudFront URL for the Control Plane SPA |
SfcControlPlaneApiUrl | API Gateway endpoint |
CognitoHostedUiDomain | Cognito Hosted UI base URL |
CognitoUserPoolId | Cognito User Pool ID |
CognitoUserPoolClientId | Cognito App Client ID |
SfcConfigBucketName | S3 bucket (configs + packages + UI assets) |
SfcLaunchPackageTableName | DynamoDB Launch Package table |
SfcControlPlaneStateTableName | DynamoDB state table (focus config) |
AgentCoreMemoryId | Short-term memory ID (also in SSM /sfc-config-agent/memory-id) |
---
Appendix — Local Development UI Setup
Note: This setup is only needed for local development against an already-deployed API. In production, the SPA is served directly from CloudFront — no local setup required.
AI-Guided Config Generation
From the Config Browser, operators can also trigger an AI-guided config creation workflow:
- Describe the machine, protocol, target AWS service, and data channels in natural language — or upload an existing spec file as context
- Optionally provide structured fields: protocol, host/port targets, sampling interval
- The agent calls the MCP server to load relevant SFC adapter and target documentation, generates a config, validates it, and saves it to S3/DynamoDB
- A job ID is returned immediately (HTTP 202); the UI polls
GET /configs/generate/{jobId}until status isCOMPLETE - The new config appears in the Config Browser, ready to be set as Focus and packaged
---
Test the SFC Config Agent (AWS CLI)
The agent runs as an Amazon Bedrock AgentCore Runtime. After deployment, retrieve the runtime ARN and invoke it:
```bash
API Gateway invoke URL (CDK output: SfcControlPlaneApiUrl)
VITE_API_BASE_URL=https://<api-id>.execute-api.<region>.amazonaws.com
Launch Packages
A Launch Package is a self-contained zip assembled by the Control Plane — everything needed to run SFC on an edge host:
launch-package-{packageId}.zip
├── sfc-config.json # SFC config with IoT credential provider injected
├── iot/ # X.509 device cert, private key, Root CA, iot-config.json
├── runner/ # aws-sfc-runtime-agent (uv / Python 3.12)
└── docker/ # Optional Dockerfile + build scriptRun on the edge host:
unzip launch-package-<id>.zip
cd runner && uv run runner.pyThe aws-sfc-runtime-agent handles IoT mTLS credential vending, SFC subprocess management, OTEL log shipping to CloudWatch, and the MQTT control channel back to the cloud.
---
Primary operator workflow
Browse Config → Edit (Monaco JSON) → Set as Focus → Create Launch Package → Download to Edge → Monitor Logs| UI Route | Purpose |
|---|---|
/ | Config File Browser |
/configs/:configId | Monaco JSON Editor — save versions, set focus, create package |
/packages | Launch Package List — live status LED, download, logs, AI-fix |
/packages/:packageId | Package Detail + Runtime Controls |
/packages/:packageId/logs | OTEL Log Viewer — ERROR-highlighted, "Fix with AI" CTA |
---
高质量的开源MCP工具，AI辅助SFC配置创建和部署
- 需要让 Claude / Cursor 操作本地工具的 AI 工程师
- 构建多智能体协作系统的 Agent 开发者
- 配置 MCP 服务器时建议使用 stdio 传输 + JSON-RPC，避免暴露公网
- Agent 任务先做 dry-run 验证工具调用链，再开启自主执行
- API key 直接提交到 git 仓库（请用 .env 并加入 .gitignore）
- MCP 配置路径拼错或权限不足，重启 Claude Desktop 才生效
- Python 依赖冲突：建议用 venv / uv 隔离环境
- 云端托管：可放在 Vercel / Railway / Fly.io 等 PaaS 平台
⚡ 核心功能
- 通过标准 MCP 协议与 Claude、Cursor 等主流 AI 客户端深度集成
- 提供结构化工具调用接口，显著降低 AI 集成复杂度
- 支持 Claude Desktop 和 Claude Code 无缝接入，开箱即用
- 可与其他 MCP 工具组合叠加，构建完整 AI 工作站
- 轻量无侵入设计，不影响现有系统架构
- 需要让 Claude / Cursor 操作本地工具的 AI 工程师
- 构建多智能体协作系统的 Agent 开发者
- 配置 MCP 服务器时建议使用 stdio 传输 + JSON-RPC，避免暴露公网
- Agent 任务先做 dry-run 验证工具调用链，再开启自主执行
- API key 直接提交到 git 仓库（请用 .env 并加入 .gitignore）
- MCP 配置路径拼错或权限不足，重启 Claude Desktop 才生效
- Python 依赖冲突：建议用 venv / uv 隔离环境
👥 适合人群
🎯 使用场景
- 在 Claude Desktop 对话中直接调用本地工具，实现 AI 与系统的深度联动
- 通过自然语言驱动复杂的多步骤自动化任务，代替繁琐手动操作
- 将多个 MCP 工具组合使用，构建个人专属 AI 工作站
⚖️ 优点与不足
- +标准化 MCP 协议，生态互联性强
- +与 Claude 官方生态无缝对接
- +即插即用，配置简单快捷
- −未明确开源协议，商用场景需谨慎评估
- −依赖 Claude 客户端，非 Claude 用户无法使用
- −MCP 协议仍在持续演进，接口可能变更
- −需要一定的配置步骤
该工具未明确声明开源协议，商业使用前请联系原作者确认授权范围，避免侵权风险。
AI Skill Hub 为第三方内容聚合平台，本页面信息基于公开数据整理，不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后，再部署至生产环境，并做好必要的安全评估。
❓ 常见问题 FAQ
经综合评估，智能SFC控制平面 在MCP工具赛道中表现稳健，质量良好。如果你已有明确的使用需求，可以直接上手体验；如果还在评估阶段，建议对比同类工具后再做决策。
| 原始名称 | sample-sfc-agentic-control-plane |
| 原始描述 | 开源MCP工具：AI assisted SFC config creation, deployment and operation.。⭐7 · Python |
| Topics | agenticaiawsmcpsfcpython |
| GitHub | https://github.com/aws-samples/sample-sfc-agentic-control-plane |
| 语言 | Python |
收录时间：2026-05-28 · 更新时间：2026-05-30 · License：未公布 · AI Skill Hub 不对第三方内容的准确性作法律背书。