ToolTrust Scanner FAQs

Question 1

What types of security vulnerabilities does it detect?

Accepted Answer

ToolTrust detects a comprehensive range of threats including prompt injection/poisoning, excessive permissions, scope mismatches, supply chain CVEs, privilege escalation, arbitrary code execution, insecure secret handling, and insufficient tool data.

Question 2

Can I integrate ToolTrust Scanner into my existing development workflow?

Accepted Answer

Yes, ToolTrust Scanner integrates seamlessly into CI/CD pipelines via GitHub Actions for automated security audits. You can also configure your AI agents (e.g., Claude Desktop/Cursor) to use ToolTrust for self-scanning external MCP servers.

Question 3

What is ToolTrust Scanner?

Accepted Answer

ToolTrust Scanner is a security tool that audits AI agent tool definitions and Model Context Protocol (MCP) servers for vulnerabilities like prompt injection, data exfiltration, and privilege escalation *before* they are executed, ensuring safe AI agent interactions.

Question 4

How does ToolTrust Scanner protect AI agents?

Accepted Answer

It intercepts and scans tool definitions prior to execution, blocking threats like prompt poisoning, excessive permissions, supply chain risks, and arbitrary code execution attempts at the source, preventing your AI agent from blindly trusting unsafe tools.

Question 5

What is the Model Context Protocol (MCP) and its relevance?

Accepted Answer

The Model Context Protocol (MCP) is a standard for defining and interacting with AI agent tools. ToolTrust Scanner is specifically designed to secure MCP servers and their exposed tools, ensuring that all AI agent interactions adhere to strict security standards.

ToolTrust Scanner

ToolTrust Scanner

Key Features

Use Cases

Key Features

Use Cases