Adversa AI March 2026 Roundup: 43% of MCP Servers Vulnerable to Command Execution, 8 Confirmed Incidents
Adversa AI·high signal
Adversa AI's March 2026 agentic security roundup documents 8 confirmed incidents across OpenClaw and ServiceNow deployments, with 43% of MCP servers found vulnerable to command execution in aggregate scanning. A new vulnerability class around persistent memory and SOUL.md identity file poisoning is emerging as the successor to prompt injection. The roundup ships a CISO playbook including detection engineering rules, incident response checklists, and 40 A2A threat papers synthesized.