Scope
Agents go where you point them.
- Targets you list
- Each scan is pointed at the hosts and repositories in its scope.
- Isolated runs
- Each run executes in its own isolated environment.
- Staging first
- Most teams start on a staging environment and widen scope once they have seen the results.
Control
You can stop it, and it stops itself.
- Hard spend caps
- Caps per scan and per project are set before anything runs. A scan stops when it’s done or when it reaches its cap.
- A stop button
- Anyone on your team with access can end a run at any point.
- Live visibility
- Every action an agent takes is recorded and visible while the run is in progress.
Models and data
You decide which models see your code.
- Model per scan
- You pick the model for each scan.
- No provider retention
- Model traffic runs through our own gateway, under provider agreements that keep none of your data.
- Covered by our DPA
- Customer data is processed under our Data Processing Agreement.
Accounts
Your identity provider, your roles.
- Single sign-on
- Sign in through your own identity provider over OIDC.
- An audit trail
- Every scan, finding and fix is kept as a record you can export.
