BitssyBytes logoAgent Safety Score

Directory / OpenAI dots

C

OpenAI dots

Always-on personal agent · OpenAI · 4/7 verified · Last checked 2026-10-16

An always-on agent with its own cloud computer. Admins decide per role whether dots can browse, reach the network, use computers or touch a local machine. By default an automatic review, not a person, decides most approvals.

Overall: C (1.50 avg, 4/7 controls verified)

Lock this down first

Leave local computer access off, and restrict app actions in Plugin controls before turning dots on.

Scorecard

CriterionRatingNotes
Permission granularity2/3Workspace owners can turn dots access and capabilities (cloud browser, network, computer use, password manager, local computer) on or off per role. Coarse on/off switches.
Human approval gates1/3Automatic approval checks apply from the start: review decides whether a dot proceeds, needs your approval or hands a step to you.
SandboxingUnverifiedNot confirmed in the docs we checked.
Secrets handling2/3Sign-in credentials go into a private form outside the conversation. Reusing a saved login needs your confirmation.
Audit logsUnverifiedNot confirmed in the docs we checked.
Kill switch / interruptUnverifiedNot confirmed in the docs we checked.
Network egress control1/3Admins can turn Cloud network access on or off. No destination allowlist is documented.

Unverified: Sandboxing, Audit logs, Kill switch / interrupt. That means we could not confirm it in public docs, not that it is missing.

Sources

Vendor? Request a correction. We review corrections within 7 days.