What we protect, and how
Access controls you set per agent
Every agent has a visibility setting that determines who can reach it and how data flows in and out. You choose the right level for each agent.
- Private: only you can see and use the agent.
- Internal: any validated member of your workspace can use it.
- External: the agent can be reached by people outside your workspace, for agents you build to face customers or the public.
Isolated, single-tenant sandboxes
Your data runs inside an isolated sandbox. Access is restricted at the network layer to validated requests, so nothing reaches your sandbox that is not supposed to. Each workspace's data lives in its own separately tenanted space.
Encrypted credential storage
Your credentials are held in dedicated secret storage and protected with standard best encryption practices, in rest and in transit.
Audit logs
Track and analyze all team member and agent activity across Skydive.
Penetration testing
Skydive conducts regular penetration testing with third-party experts. We also have automated static analysis checks and dependency scanning across our codebase and agent codebases.
Compliance
Skydive is SOC 2 Type I certified. We also comply with the Google API Services User Data Policy and have passed a Cloud Application Security Assessment (CASA) Assurance Level 1 (AL1) audit.
