All insights

Building a multi-agent security pipeline with Google's A2A protocol and GitGuardian

Andy Rea, our co-founder, published a guide on the GitGuardian blog showing how to build a multi-agent security pipeline using Google's Agent-to-Agent (A2A) protocol.

What the demo shows

The demo wires up multiple AI agents using Google's Agent Development Kit (ADK) and the A2A protocol. The pipeline integrates GitGuardian to scan content for secrets, which makes it a practical example of multi-agent orchestration doing security work rather than describing it.

Key technologies

  • Google Agent Development Kit (ADK) — Google's framework for building and deploying AI agents
  • A2A protocol — Google's Agent-to-Agent protocol for inter-agent communication
  • GitGuardian — secrets detection and security scanning

Why this matters

As agents take on more of the development workflow, the ability to orchestrate several specialised agents safely becomes the constraint. This build is a blueprint for security-focused agent pipelines that catch secrets before they leave the room, and it is the kind of work that sits behind our agentic AI practice: run it ourselves first, then advise on it.

Read the full article on GitGuardian

Contact

Talk to us before your next AI investment.

We run agents in our own business every day. If you are planning to run them in yours, the first conversation is diagnostic, not a pitch.