UseAgentSECURITY
Keep control of agent work.
Choose what your agents can access, review sensitive actions, and follow the work through its recorded history.
Access with a defined scope
Short-lived capability tokens bind supported operations to a particular organization, user, run and permission scope.
Separate workspaces for agents
Agents work in isolated Linux environments, with supported tool access routed through the gateway. Your deployment and provider choices determine where work runs.
A person reviews sensitive actions
Approval flows show the requested operation before it proceeds. Your decision is recorded with the run, so the team can see what was approved.
Protected credential storage
Stored organization secrets are encrypted and sensitive values are redacted from response logs. Review your provider and integration configuration when setting up a workspace.
A history you can inspect
Follow the sequence from request and tool calls to approvals and results. Saved run history lets the team review what happened later.
Organization-aware access
Organization scoping and identity checks control access to team resources. Administrators choose integrations and manage who can work in the workspace.
Start with one task.
See what you can hand off.
Explore a completed run, or join the early-access list to hear about getting your team started.