Permissions
Checks whether the agent has more access than its task requires.
Free AI security check
MCP Server Risk Assessment
Review tool trust, server permissions, dynamic discovery, secrets, supply-chain exposure and action validation.
Example residual risk46/100
Moderate risk
What the assessment covers
Untrusted input
Reviews how external content can influence tools and actions.
Autonomy
Measures approval gates, limits and potential blast radius.
Evidence
Evaluates logging, testing and incident containment.