Radware expanded its Agentic AI Protection product on July 7 with three capabilities aimed at a gap most agent security vendors have ignored: AI agents that run on developer laptops instead of cloud environments.
The release adds real-time monitoring for developer-hosted agents like Anthropic’s Claude Code, audit-ready compliance reporting mapped to ISO 42001, the EU AI Act, and NIST’s AI Risk Management Framework, and deeper visibility tools that chart how agents interact with enterprise applications, services, and resources.
The Developer Endpoint Problem
Most agent security products assume agents run as SaaS services in cloud environments. Claude Code does not. It runs directly on a developer’s machine, accessing local files, executing shell commands, and interacting with codebases through the terminal. That local execution model puts it outside the reach of cloud-based security monitoring.
Radware’s update extends protection to these local-host endpoints, according to SiliconANGLE. Security teams can now monitor agent behavior across conversations, govern which tools an agent is allowed to access, and apply sensitive-data guards on developer machines.
“Organizations are deploying AI agents across increasingly complex environments, creating new requirements for visibility, governance and security,” David Aviv, Radware’s chief technology officer, told SiliconANGLE. “These enhancements help organizations better understand agent behavior, support their compliance efforts and help extend protection to AI agents operating across both SaaS and local developer-hosted environments.”
Compliance Pressure Driving Demand
The compliance reporting component reflects growing regulatory pressure on enterprises deploying autonomous AI. Three frameworks now dominate the conversation: ISO 42001 (the international standard for AI management systems), the EU AI Act (which classifies certain agent uses as high-risk), and NIST’s AI Risk Management Framework (the US federal guidance for AI governance).
Radware built its reporting to produce audit-ready documentation showing how organizations track, contain, and govern the risks tied to AI agent activity. For enterprises preparing for EU AI Act enforcement or seeking ISO 42001 certification, that paper trail is increasingly a procurement requirement.
Agent Visibility at Scale
The visibility improvements add monitoring and mapping features that chart agent activity, interactions, and dependencies across the enterprise. As agent deployments multiply, security teams have struggled to see how autonomous systems connect to applications, services, and other resources. Radware’s update addresses that with dependency mapping that tracks agent-to-service connections in real time, according to the company’s press release.
The Broader Security Shift
Radware, a NASDAQ-listed company (RDWR) historically known for DDoS mitigation and application security, has been expanding into AI security as enterprise agent adoption accelerates. The company partnered with Dataiku in June 2026 to integrate its security capabilities into Dataiku’s AI platform.
The new capabilities are available immediately as part of the existing Agentic AI Protection product.