AI Agent Sandbox Risk Checker
Map agent capabilities to sandbox review needs. This result is a practical review aid, not a live test or formal security audit.
TOOL_RC_01
Risk Assessment Entry
Describe capabilities, not sensitive implementation details. The rules engine will return your review checklist.
Access scope
No secrets. Do not submit API keys, credentials, source code, private logs, or customer data.
Your agent risk profile
Your risk checklist will appear here after you describe your agent and run the check.