You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Decide semantics: are we asking "Are guardrails present?" or "Is the condition present?"
alternative suggestion: insufficiently mitigated, sufficiently mitigated makes it less ambiguous for each, but could be difficult to ascertain
Warning
The presence or absence of guardrails does not say anything about their sufficiency to any particular attack. So it's not clear how useful this information will be without additional refinement. See #1054 for any potential resolution of this tension.
Desired State
Add one new decision point for to represent each of the 10 Core Agentic AI Risk Items
For each item, values will be binary:
Warning
The presence or absence of guardrails does not say anything about their sufficiency to any particular attack. So it's not clear how useful this information will be without additional refinement. See #1054 for any potential resolution of this tension.