Snyk Labs

As organizations adopt AI agents, visibility into what those agents can actually do, including tools, permissions, external dependencies, and hidden execution paths, becomes increasingly critical. Traditional security tooling was never built to understand agent “skills” or MCP-based supply chains.

With AI agent skills becoming part of real developer workflows. A single install command can connect an autonomous agent to your codebase, system tools, or external services. Unlike installing a passive library, these skills can execute logic, retrieve data, and trigger downstream actions on your behalf.

Snyk Agent Scan - Skill Inspector helps developers and security teams detect malicious skills, insecure configurations, and leaked secrets before they reach production, and until now, it has been available via our CLI.