Loading connector details…
Loading connector details…
Choose a unique username to continue using AgentHotspot
by sattyamjjain • Uncategorized
An open-source security firewall decorator for AI agents that enforces strict type validation, argument stripping, sandboxing, and policy controls.
Safely execute tool calls with strict input validation and argument sanitization.
That require isolated sandboxed execution environments to prevent harmful system operations.
Enforce security policies, cost controls, and observability during tool execution.
Agent-Airlock provides a single decorator to secure AI agent tool calls by stripping invented arguments, enforcing strict types, and sandboxing execution in isolated microVMs. It offers features like role-based access control, PII masking, network airgap, cost control, and observability with OpenTelemetry. This open-source solution prevents harmful or unintended actions by AI agents, ensuring safe and controlled execution without vendor lock-in.