level-4: Naman Anand

[Naman-Playz]

Submission Level

Level: level-4

What I Did

I have completed the Level 4 challenge by building a Secure Agent Mesh that facilitates collaboration between two specialized AI agents.

• A2A Discovery: Implemented two agents—the SMILE Security Auditor and the Infrastructure Architect that discover each other using standards-compliant A2A Agent Cards.
• Secure Infrastructure Blueprint: The agents communicate via structured JSON messages (HMAC-signed) to combine LPI knowledge into a Hardened Home Network Digital Twin Blueprint a value-add output the LPI cannot generate alone.
• Security Hardening: The mesh is natively hardened against four classes of attack: Prompt Injection, Data Exfiltration, Denial of Service (DoS), and Privilege Escalation.

Checklist

• I read the README and CONTRIBUTING guide
• My PR title follows the format: level-4: Naman Anand
• I tested my changes locally before submitting
• Threat model and security audit included
• Structured data exchange verified via HMAC signatures

Signed-off-by: Naman Anand ns398592@gmail.com