GitHub - vellaveto/vellaveto: Runtime security engine for AI agent tool calls. Policy firewall for MCP with <5ms P99, 20+ threat detection layers, formal verification. One command: vellaveto-proxy –protect shield · GitHub

The vellaveto GitHub post introduces VellaVeto, a runtime security engine designed for AI agent tool calls, which intercepts requests, enforces policies, and maintains an audit trail. It offers various deployment modes, including a Consumer Shield for user-side privacy protection, and emphasizes robust security with extensive testing, formal verification, and compliance with multiple regulatory frameworks.

Edward Kiledjian @ekiledjian