Snyk Agent Scan
OfficialScans AI agents, MCP servers, and agent skills for security vulnerabilities including prompt injections, malware payloads, and credential issues. Auto-discovers components from Claude, Cursor, Windsurf, and other agent platforms.
1,779208 views181Local (stdio)
What it does
- Auto-discover MCP configurations and agent tools
- Scan for prompt injection vulnerabilities
- Detect malware payloads in natural language
- Check for credential handling issues
- Identify tool poisoning and shadowing attacks
- Scan Claude, Cursor, Windsurf, and Gemini CLI setups
Best for
AI developers using MCP serversSecurity teams auditing agent deploymentsOrganizations using AI coding assistantsDevelopers building custom agent skills
15+ distinct security risk types1,700+ GitHub starsTechnical report on agent ecosystem threats
