Respond to risk recommendations related to AI agents to close the security gaps
before bad actors can exploit them.
| Where Can I Use This? | What Do I Need? |
|
|
Or any of the following licenses that include the SaaS Agent Security license:
- CASB-X
- CASB-PA
- SaaS Security Posture Management license
|
SaaS Agent Security detects agent risks and recommends remediation actions
you can take to close the security gaps before bad actors can exploit them. For
example, SaaS Agent Security can detect the following risks and suggest
remediation actions.
- Agents with access to sensitive knowledge bases, which could result in
unauthorized data exposure.
- Agents that have not shown any activity for an extended period. Because these
dormant agents are not being actively managed, they could represent a hidden
attack surface. Their inactivity might delay detection of malicious use.
- Agents that have elevated, administrator-level access to the applications that
they connect to, which might be a violation of the principle of least privilege
(PoLP).
- Agents without delegated permissions. By interacting with the agent, users might
indirectly gain access to information and actions that they are not directly
authorized to access.
For a complete list of the risks that
SaaS Agent Security can detect, refer
to
Risk Recommendations.