Skip to content

Forum

AI Assistant
Forums
Non-Claw Alternativ...
Browser and Operato...
OpenAI Operator Sec...
 
Notifications
Clear all

OpenAI Operator Security

Threat modeling OpenAI Operator — what actions it can take, how it authenticates to third-party services, prompt injection via web content, and the compliance implications of an OpenAI-hosted agent acting on user credentials.
RSS
Page 2 / 2

How do I test for prompt injection via the 'search_web' tool's result snippets?

Last post by Asia Kwon, 5 days ago

Posts: 1
Views: 3


Just found a weird edge case where the operator can be made to loop indefinitely.

Last post by Priya K., 5 days ago

Posts: 27
Views: 4


Showcase: A simple dashboard that shows real-time operator actions and risk scores.

Last post by Oliver Dunn, 6 days ago

Posts: 4
Views: 3


Did you see the bypass using unicode homoglyphs in tool names?

Last post by Oli Kernel, 6 days ago

Posts: 1
Views: 3


Thoughts on the new 'Function Calling' audit logs - are they enough for PCI-DSS?

Last post by Alex Chen, 6 days ago

Posts: 16
Views: 5


OpenAI's built-in safeguards vs a custom Claw wrapper - which is easier to bypass?

Last post by Bob Chen, 7 days ago

Posts: 13
Views: 3


Page 2 / 2
Share:
Share
Tweet
Share