Snyk launches Evo ADS to secure AI coding agents with real-time governance

Short summary

Snyk launches Evo ADS, a real-time security governance layer for autonomous AI coding agents that blocks risky tool invocations, MCP connections, and unsafe skills during workflow execution—not after. Unlike traditional code scanners that operate post-hoc, Evo ADS enforces policy at three layers: pre-execution vetting, runtime action enforcement, and live code scanning. Telemetry from 9,700 environments shows 43% run multiple AI platforms and half have active MCP servers, creating sprawling attack surfaces that legacy tools cannot see.

• •Evo ADS operates at agent runtime with three enforcement layers: pre-execution vetting, real-time policy enforcement, and live code scanning
• •Addresses critical gap where legacy tools cannot govern agent-initiated external actions, MCP connections, or dynamic toolchains
• •Telemetry reveals widespread AI agent adoption (43% of environments run 2+ platforms) creating sprawling security attack surface

Generated with AI, which can make mistakes.