Helmet Security Expands Platform to Secure the Broader Agentic AI Ecosystem
New browser-based controls, expanded integrations and deeper activity visibility extend agentic AI discovery, monitoring and policy enforcement across enterprise environments
Helmet Security has significantly expanded its agentic AI security platform, tripling its integration ecosystem and introducing new capabilities for agent discovery, activity monitoring and policy enforcement. The updates extend Helmet’s reach across the enterprise, including AI activity that never touches an endpoint, while enabling organizations to build on their existing security investments.
The expansion reflects Helmet’s approach to securing a rapidly evolving AI ecosystem. The company continuously adapts its platform to address emerging technologies, frameworks, and standards, so customers don’t have to independently navigate every new development.
What’s New: Expanded Agentic AI Security Capabilities
Helmet’s latest platform updates include:
- Browser-Based Policy Enforcement: Now generally available, Helmet’s browser extension provides an additional deployment option for enforcing policies and allowlisting approved AI agents, MCP servers and skills. It also enables organizations to block AI functionality within everyday applications like Gmail, Google Drive and the AWS Console without disrupting the rest of the site.
- AI Agent Kill Switch: Instantly pause or terminate a misbehaving AI agent, automatically or with one click, through Helmet or the customer’s existing EDR.
- Intent Guard: Blocks agent actions that don’t match what the user actually asked for.
- Expanded Enterprise Integrations: Helmet has tripled its integration ecosystem, connecting with technologies across endpoint, identity, device management, cloud and SaaS environments. Integrations include identity providers, endpoint detection and response providers, as well as hyperscalers, enabling organizations to extend AI policy enforcement through their existing security investments.
- Deeper Agentic Activity Visibility: API-based connectors extend discovery to AI operating beyond endpoints. Helmet’s Anthropic Compliance connector analyzes conversation-level activity to identify the MCP server behind individual tool calls, providing context beyond application usage or summary activity feeds.
- AI Supply Chain Protection: Detects poisoned or tampered MCP servers, including those that modify their available tools after receiving approval, helping organizations identify risks introduced through the AI supply chain.
- Local AI Model Discovery: Expands visibility to models running locally on endpoints, including Ollama and LM Studio, alongside AI agents, MCP servers and skills.
- Enhanced Policy Enforcement: Expanded policy capabilities give security teams greater flexibility to establish and apply controls across AI agents, MCP servers and associated tools. New inference hooks enable platforms like Anthropic and Glean to obtain Helmet’s approval before prompts execute, extending enforcement to environments without endpoint involvement.
Keeping Pace With an AI Ecosystem That Won’t Stand Still
According to EY’s AI governance research, 69% of senior AI decision-makers are concerned their organizations lack the internal expertise needed to continuously evolve AI governance controls. Among organizations using agentic AI, 49% have yet to update their governance frameworks specifically for agentic AI risks and requirements.
These findings underscore a growing responsibility for security vendors. As AI infrastructure, protocols and frameworks evolve, organizations need security technology that keeps pace without requiring their internal teams to become experts in every new development. Helmet’s rapid development cycle is designed to address that challenge, translating changes across the AI ecosystem into new integrations, visibility and controls while making existing security tools more capable.
“Security teams shouldn’t have to become experts in every new AI framework or protocol just to maintain control of their environments,” said Kaushik Shanadi, CTO of Helmet Security. “That’s why we’re constantly evolving Helmet alongside the technologies our customers adopt, while helping them get more from the security infrastructure they already have. AI should make those existing tools smarter, and our job is to provide the context needed to understand and secure agentic behavior.”