Agent Card Prompt Injection: The Security Nightmare of AI Agent Discovery

How malicious agents can hijack A2A protocol discovery through prompt injection in Agent Cards — and why defense-in-depth is the only real answer.