Search K
Appearance
July 28, 2026 - SaaS - v10.81.3
INFO
AI Guardrails now supports scanning responses.
This release includes updates in the following areas for F5 AI Security:
AI Guardrails
Scanning responses
F5 guardrails scan prompts by default, preventing PII and unauthorized requests from reaching the model. You can now also scan responses and prevent unauthorized content from leaving the model and reaching end-users or agents.
For best results with this feature, we recommend you:
- Enable response scanning only where needed. Enabling both prompt and response scanning increases latency and token use. Responses are often much longer than the original prompt.
- Test before deploying. F5's guardrails efficacy data applies to prompts only.
AI Red Team
New agentic resistance attack
We released a new multi-turn agentic resistance attack called echo chamber. The echo chamber attack plants "poisonous seeds," innocuous sentences that contain keywords related to a harmful objective. Over multiple turns, it manipulates the model into developing those ideas, gradually steering the conversation toward harmful content.
Agentic resistance attacks use a model to replicate real adversarial techniques. The model formulates attacks and analyzes responses in real-time. It evolves its approach just as a human attacker would, including backtracking when the attack isn't working.
Attack pack updates
The August 2026 attack pack includes 10,000+ novel attack prompts and a new attack vector: morality dilemma. This attack manipulates the model's ethical alignment by embedding misleading or competing moral norms into the prompt.
Improved runtime
Process optimizations reduce total runtime for red-team campaigns by up to 10%.
Resolved issues
This release fixes the following issues:
- Raw data view for operational attacks now includes full details.
- Custom guardrail message now shows number of affected projects.
- Outcome analysis now highlights correct text in JSON response.
- Audit log filters now update the URL without a full page reload.
Doc updates
Since the last SaaS release on July 8, AI Security has added the following documents:
INFO
The table of contents has been extensively changed. Content is now organized by function, such as Get started with AI Security, AI Guardrails, AI Red Team, and About the API.