Research explained
Radware expands its AI security offering beyond model-input filtering
The February 2026 product coverage describes vendor capabilities, rather than an independently validated guarantee of agent security.
What happened
MSSP Alert reported that Radware expanded its AI security portfolio with agent-focused protection alongside its LLM Firewall. The offering targets enterprises and service providers deploying connected AI applications.
This is product reporting based partly on vendor statements, not a vulnerability disclosure or proof that a firewall can eliminate prompt injection. Buyers should distinguish model-content filtering from controls over tool permissions and application actions.
Evaluate proposed products against realistic workflows, including private data access, connector behavior and failure handling. Ask what is enforced outside the model and which events remain visible to the organization’s security team.
Next step
Test proposed controls against connected-agent workflows and observable failure cases.
What remains unknown
- Open question
- The sources reviewed do not establish whether your deployment was exploited.
Source references
-
MSS