Integration brings real-time security evaluation to LLM & MCP interactions without requiring changes to agents or developer workflows

SAN FRANCISCO: AI agents now generate thousands of LLM requests across developer machines, CI pipelines, and internal systems. Each request can carry prompts, tool calls, secrets, and sensitive data. This creates a new and largely unmonitored security surface for enterprises. Highflame, an AI Security company focused on securing AI agents and their interactions, today announced a partnership with Tailscale to bring real-time security evaluation to AI activity at the network layer.

By integrating Aperture by Tailscale with Highflame’s platform, organizations gain visibility into LLM interactions and the ability to assess risk across prompts, tool usage, and model outputs. This works without requiring changes to agents or developer workflows.

Aperture by Tailscale provides a centralized gateway for AI traffic. It routes requests through the network and captures usage, identity, and telemetry. Highflame builds on this by analyzing each interaction to detect risks, including prompt injection, secret/credentials/, and PII leakage, unsafe tool execution, and policy violations.

“AI agents are already operating across every layer of the enterprise, but security hasn’t caught up to where the activity actually happens,” said Sharath Rajasekar, CEO of Highflame. “Partnering with Tailscale allows us to secure every AI interaction, without requiring developers to change how they work.”

“Aperture gives organizations a reliable control point for AI traffic,” said Avery Pennarun, CEO of Tailscale. “With Highflame, customers can take that further by understanding the security implications across prompts, tool calls, and model responses, turning visibility into something they can actually use.”

Unified Visibility and Control Across AI Systems

The integration creates a unified layer of visibility and control across both the agent and network layers. This allows organizations to continuously evaluate AI activity, enforce security policies, and maintain visibility into how AI systems operate.

The solution runs in the background and does not require SDKs or instrumentation. Developers can continue using their preferred tools. Security and compliance teams gain centralized visibility, along with detailed logs that include identity, context, and policy outcomes.

Organizations using Aperture by Tailscale can enable the Highflame integration with minimal configuration. Aperture is currently in alpha and available to early users.

Learn more: Visit https://highflame.com/code-agent-control-plane | https://tailscale.com/use-cases/securing-ai

About Highflame

Highflame is an Agent Security and AI Governance company building the runtime control plane for agentic AI systems. Its platform enables organizations to establish agent identity and monitor, evaluate, and enforce security policies across AI interactions in real time. Highflame is headquartered in the San Francisco Bay Area and backed by leading investors.

For more information, visit https://highflame.com.

About Tailscale

Tailscale provides secure, identity-first networking, simplifying complex network setups with fast, reliable connections that seamlessly scale across cloud and on-premises environments. Trusted globally by over 25,000 businesses, including Microsoft, NVIDIA, Telus, and Instacart, and millions of individuals, Tailscale’s award winning technology enables effortless adoption of Zero Trust principles with secure connectivity that just works.

Fonte: Business Wire