- Agent Infrastructure: Start Here
- Agent Interop with A2A & MCP
- Sandboxed Agent Execution
Ship production agents that call other agents and run untrusted code safely, using A2A, MCP, signed agent cards, and E2B/Modal/Daytona sandboxes. 3 advanced chapters for engineers.

Info basate su dati pubblici. I dati potrebbero essere imprecisi o non aggiornati: verificali sul sito del provider. Formazione non è responsabile del corso, organizzato dal provider, che è il solo responsabile dell’iscrizione e dell’erogazione del corso.
Corso da seguire online, organizzato da Anthropos AI Academy.
Ship production agents that call other agents and run untrusted code safely, using A2A, MCP, signed agent cards, and E2B/Modal/Daytona sandboxes. 3 advanced chapters for engineers.
It is built for software engineers who already work with AI agents and need to move them from prototype to production. The material is advanced and assumes you write code and understand agent basics.
Connect agents across frameworks using the A2A protocol and Model Context Protocol (MCP)
Issue and verify signed agent cards for secure agent-to-agent delegation
Bridge agents built in different frameworks so they exchange tasks
Choose between E2B, Modal, and Daytona isolation primitives for a given workload
Run untrusted, agent-generated code inside a sandbox without exposing the host
Design the runtime substrate, memory backends, and orchestration that production agents depend on
You learn to build agents that communicate over the A2A protocol and MCP, secure delegation with signed agent cards, and run untrusted agent-generated code inside E2B, Modal, or Daytona sandboxes. It covers the runtime substrate, interop protocols, and isolation choices that production agents rely on.