
Claude Agents
Problem
Most AI agents are glorified prompts - they respond, they don't act. Real autonomous workflows that connect external systems, execute multi-step tasks, and stay auditable require a proper technical foundation: MCP, Skills, Subagents. The reality: that's not something you pick up from a 20-minute YouTube tutorial.
Solution
Model Context Protocol (MCP) is the open standard that lets Claude communicate securely with external tools and data sources - without writing tool schemas by hand. We combine the official Anthropic MCP curriculum with practical agent design patterns and DACH compliance context. The goal: participants leave with a working MCP server and the full toolkit for production-ready agents.
Agenda
Execution
Basic Python, async/await understanding, and API familiarity required. Max 8 participants per cohort.
Compact (3h): Modules 01-02. First MCP server running, architecture understood.
Full Day (8h): All four modules + complete agent implementation + governance review.
Language: English or German.


Target Audience
Software & platform engineers: Who want to connect Claude to internal systems.
ML/AI engineers: Building production-grade agent workflows.
Tech leads: Who want a solid technical foundation for their organisation's agent strategy.
ROI & Business Impact
Why This Training Is Essential
No More Boilerplate
MCP eliminates hours of integration code. Once you understand it, implementation is dramatically faster.
Real Autonomy
Agents that connect external tools, execute multi-step tasks, and remain auditable - not just prompts with a UI.
Future-Proof
MCP is the open standard. What you learn here works with Claude today and with the next model tomorrow.
Pricing
Simple, Transparent Investment
FAQ
Common Questions
Do I need to know Python?
What is the difference from AI Agents & Secure Automation?
Are there ready-made MCP servers we can use?
Is BAFA funding available?
That's not all








