Search K
Appearance
September 30, 2026 - SaaS - v10.122.0
This release includes updates in the following areas for F5 AI Security:
AI Red Team
Attack vector and intent matrix
Red Team reports now includes a matrix which shows attack success rates, dependent on attack vector and intent category.
Attacks by framework
Red Team reports now includes attack success rates for two criteria associated with the OWASP LLM Top 10 for 2026:
- Prompt injection
- Misinformation
AI Guardrails
Arabic language integration and PII package
PII packages now include Arabic language PII support for two countries: Saudi Arabia and the United Arab Emerates. In the UI, see AI Guardrails > Guardrails and look for available PII packages.
For on-premises installations, you must reinstall AI Security by redeploying to the cloud or using Helm.
Scan summary metrics
The AI Security dashboard now shows the overall volume of scanned traffic. Use these metrics to review the quantity of scans, prompts, and total traffic:
- Scans: Number of scan requests made. Also see the Scans API.
- Prompts: Number of prompts made. Also see the Prompts API.
- Blocked Prompts: Number of blocked prompts.
- Total Traffic: Total number of requests and responses scanned.
Known issues
AI Red Team report, print report view, does not contain charts including:
- OWASP 2026 attack success rates
- Attack success by attack vector
Resolved issues
This release fixes the following issues:
- A project API token could grant users read access to information they should not see.
- A failed provider scan could return a CASI score of 100%.
- Search functionality did not work when opening the Import and deploy dropdown in Remediation.
- Permissions for Remediation access did not work when assigned from the Remediations > Remediation access page.
Doc updates
Breaking changes
There are no breaking changes or required customer actions in this release.