Amine Raji, PhD
Articles sur Amine Raji, PhD
Why Your CVE Program Misses Most AI Agent Vulnerabilities
CVEs still name code bugs well. For AI agents and MCP servers, most real risk sits in defaults, prompt injection, tool composition and supply chain, where a CVE says little or nothing. A map of where it helps, where it misleads, and the triage, inventory and controls that replace it.
The Lethal Trifecta in Production: How Enterprise Copilots Quietly Became Exfiltration Tools
Across 2025, enterprise copilots with private data access, untrusted content exposure, and outbound paths quietly became data exfiltration tools. Here is the pattern and the fix.
The Trust Gap: Your Threat Model Stops at the Protocol Boundary
MCP security analysis does not consider A2A routing. A2A security analysis does not consider MCP poisoning. The attack lives in the gap between two layers, and every tool in the coverage matrix evaluates one protocol at a time.