SkillAgentSearch skills...

mcp-hack26-tia-connect-agentgateway

MCP_HACK//26 — Industrial Zero-Trust AI: T-IA Connect + AgentGateway for Siemens TIA Portal

Install / Use

claude mcp add feelautom -- npx -y github:feelautom/mcp-hack26-tia-connect-agentgateway

If the server publishes to npm under a different name, use that package instead — check the repo README.

About this skill
🔌

MCP Server

Model Context Protocol server

Quality Score

72/100

Category

Security

Supported Platforms

Claude Code
Claude Desktop

Our assessment of mcp-hack26-tia-connect-agentgateway

mcp-hack26-tia-connect-agentgateway scores 72/100 on our quality scale, 1034th of 1,100 Security skills we index.

Its MCP Server is 5.5 KB long, well organised into 20 sections with 5 code examples: a solid amount of guidance for an agent.

It has 10 GitHub stars, so there is little community track record yet; judge it on its content.

Substance
26/30
Structure
20/20
Description
12/15
Adoption
4/20
Freshness
11/15

Maintenance, license and trust

  • The repository was last updated about 8 months ago. That is recent enough to be usable, but agent tooling moves fast, so check the instructions against your agent's current version.
  • Our last check on 2026-10-22 found the source still online.
  • No license is declared. By default that means all rights are reserved: you can read it, but reusing or redistributing it is not clearly permitted. Ask the author before building on it commercially.
  • Its trust signals score 74/100, with 3 cautions from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.

Safety scan

No issues found

Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.

AI review by kimi-k2.7-code on 2026-09-24. Automated pattern scan on 2026-09-24. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.

mcp-hack26-tia-connect-agentgateway compared with similar skills

All 4 of these similar skills score higher than mcp-hack26-tia-connect-agentgateway; compare them before choosing.

SkillScoreStarsUpdatedFormat
mcp-hack26-tia-connect-agentgateway (this skill)by feelautom72108mo agoMCP Server
Agent-Reachby Panniantong100110.0ktodayCLAUDE.md
headroomby headroomlabs-ai10079.1ktodayCLAUDE.md
CowAgentby zhayujie10047.8k1d agoCLAUDE.md
Scraplingby D4Vinci10099.4k2d agoMCP Server

Frequently asked questions

How do I install mcp-hack26-tia-connect-agentgateway?
Run claude mcp add feelautom -- npx -y github:feelautom/mcp-hack26-tia-connect-agentgateway. The install tabs above show the steps for each supported agent.
Which AI agents does mcp-hack26-tia-connect-agentgateway work with?
It is written for Claude Code and Claude Desktop, as a MCP Server file. Other agents that read the same format can often use it too.
Is mcp-hack26-tia-connect-agentgateway safe to use?
Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It declares no license and scores 74/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
Is mcp-hack26-tia-connect-agentgateway still maintained?
The repository was last updated about 8 months ago. That is recent enough to be usable, but agent tooling moves fast, so check the instructions against your agent's current version.

T-IA Connect + AgentGateway — Industrial Zero-Trust MCP

MCP_HACK//26 Submission — Solo.io & a16z Hackathon

AI-powered TIA Portal with an industrial-grade Zero-Trust proxy. T-IA Connect exposes Siemens TIA Portal as an MCP server — AgentGateway secures every single call.

Demo Video

Watch the demo

Blog Post

T-IA Connect + AgentGateway: Industrial Zero-Trust AI in Action


Quick Overview for Judges

What: An AI agent (Claude) that creates, configures and manages Siemens TIA Portal PLC projects — with every MCP call routed through AgentGateway for Zero-Trust security.

Why: In industrial automation, AI agents interacting with PLCs need strict guardrails. A single unauthorized command (stopping a CPU, closing a project) can halt a production line. AgentGateway adds the missing security layer.

How:

Claude Code (CLI) → AgentGateway (port 4000) → T-IA Connect (port 9000 SSE)
                                              → T-IA Connect REST (OpenAPI → MCP)

Key Results:

  • 190+ MCP tools exposed from TIA Portal (SSE + OpenAPI merged)
  • Zero-Trust policy enforcement — blocked tools disappear from the agent's view
  • Full audit logging of every AI action
  • Real-time demo: project creation, CPU config, ladder logic — all via natural language

Architecture

Components

| Component | Role | Port | |-----------|------|------| | T-IA Connect | MCP Server for Siemens TIA Portal V20 (Openness API) | 9000 | | AgentGateway | Zero-Trust MCP proxy (Solo.io) | 4000 | | Admin / Playground | AgentGateway UI | 15000 | | Claude Code | AI Agent (Anthropic CLI) | — |

Two Backends, One Endpoint

AgentGateway merges two backends into a single MCP endpoint:

  • siemens-tia-portal_* — Native SSE tools from T-IA Connect
  • tia-openapi_* — Full REST API auto-converted to MCP tools via OpenAPI spec

Deployment

Prerequisites

  • Windows 10/11 with Docker Desktop
  • Siemens TIA Portal V20 (with Openness API enabled)
  • T-IA Connect v2.0.335+ running on port 9000
  • Claude Code CLI

1. AgentGateway Configuration

See agent-gateway.yaml for the full configuration.

Key features:

  • SSE backend pointing to host.docker.internal:9000 (Docker→Windows bridge)
  • OpenAPI backend with recursive-ref-free spec (tia-openapi3-norec.json)
  • CORS configured for Playground UI
  • Host: localhost:9000 header override (required by Windows HTTP.sys)

2. Start AgentGateway

docker run -d --name agent-gateway --restart always \
  -p 4000:4000 -p 15000:15000 \
  -e ADMIN_ADDR=0.0.0.0:15000 \
  -v "$(pwd)/agent-gateway.yaml:/config.yaml" \
  -v "$(pwd)/tia-openapi3-norec.json:/tia-openapi3-norec.json" \
  ghcr.io/agentgateway/agentgateway:v1.0.0-alpha.2 -f /config.yaml

Important: Use v1.0.0-alpha.2 specifically. The latest tag causes stack overflow on OpenAPI specs with recursive references.

3. Configure Claude Code

In your .claude.json, add the MCP server:

{
  "mcpServers": {
    "tia-connect": {
      "type": "http",
      "url": "http://localhost:4000/mcp"
    }
  }
}

4. Verify

  • Open AgentGateway Playground: http://localhost:15000/ui
  • Check that both backends are connected
  • Call tia-openapi_Health_Ping to verify TIA Portal connectivity

Zero-Trust Policy Example

Block the close_project tool so no AI agent can close a TIA Portal project:

policies:
  mcpAuthorization:
    rules:
    - deny: mcp.tool.name == "close_project"

Once applied, the tool disappears from the agent's tool list entirely. No error, no bypass — it simply doesn't exist from the agent's perspective.


OpenAPI Spec Processing

The TIA Portal REST API spec contains recursive $ref references that cause stack overflow in AgentGateway. We provide a script to remove them:

node fix-recursive-refs.js
# Input: tia-openapi3-slim.json (622KB, 146 paths)
# Output: tia-openapi3-norec.json (368KB, no recursive refs)

See fix-recursive-refs.js for details.

Related AgentGateway issues: #132, #224


Known Limitations

  • Hot reload not supported: Policy changes require a container restart (alpha software)
  • OpenAPI recursive refs: Must be pre-processed with fix-recursive-refs.js
  • Windows only: T-IA Connect requires Windows + TIA Portal (Openness API is Windows-native)

Tech Stack

  • Siemens TIA Portal V20 — Openness API
  • T-IA Connect v2.0.335 — Industrial MCP Server (t-ia-connect.com)
  • AgentGateway v1.0.0-alpha.2 — Solo.io
  • Claude Code — Anthropic CLI

Links


Author: Franck Pichard — feelautom.fr

Related Skills

View on GitHub
GitHub Stars10
CategorySecurity
Updated7mo ago
Forks1

Languages

JavaScript

Trust signals

74/100

From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.

1 medium2 low1 info