Navigating the EVO Interface

Before you jump into the live environment, it's important to understand that EVO isn't just a static dashboard, it's an agentic security partner. While traditional Snyk tools focus on finding vulnerabilities in static code, EVO focuses on unifying security across the agent lifecycle, by bridging development and runtime with continuous enforcement and visibility. Secure both agentic development to highlight what agents use, do, and generate and how agentic applications run, what they access, and what they impact.

To navigate EVO effectively, you need to understand the "Agents" working behind the scenes. These agents function as dedicated sub-processes, each managing a distinct phase of the AI security lifecycle.:

• Discovery Agent (The "Eyes"): Constantly maps your AI inventory. It identifies not just code, but the models, datasets, and Model Context Protocol (MCP) servers your applications are actually using.
• Risk Intelligence Agent (The "Brain"): This agent doesn't just look for CVEs; it evaluates the "behavioral" risk of AI models (e.g., bias, propensity for hallucinations, or safety guardrail bypass) based on Snyk's proprietary risk index.
• Policy Agent (The "Guardrail"): This agent takes your high-level business rules via natural language prompts (for example, "No unencrypted PII in LLM prompts") and enforces them across your entire environment.

In EVO, there are two ways to work. There is no "right" way; EVO is designed to meet you where you are:

• The Workflow Agent (Natural Language):

  • Located on the right side of the interface. This is conversational interface. If you need to see your "attack surface," you don't need to build a query, you just ask by entering a prompt.
  • NOTE: EVO will often suggest in context, personalized "Follow-up Prompts" after it answers your initial query. These are designed to help you dig deeper into a risk you might have missed.

• The Inventory & Policy Tabs (Manual Navigation)

  • For those who prefer a traditional "Point-and-Click" experience, the sidebar provides direct access to specific repositories to see exactly which models are being called and what issues are open.

In this demo, you will learn the elements of the interface by playing the role of a Security Engineer performing a health check:

• Interact with the Workflow Agent to get an overview of your AI attack surface.
• Explore the AI-BOM to see how EVO indeitifies "Shadow AI" (untracked models).
• Validate a Risk Score for a specific model to understand why it was flagged.
• Save your chat results as a live-updating Report.

Ready to perform a health check? Click "Start Demo" below.