Reco Adds Runtime Security Controls for AI Agents

Reco has expanded AI Runtime with Browser Guard, prompt and tool call analysis, and Smart Remediation, detailed in a company blog post.

Browser Guard is designed to show which employees are using AI tools in browsers, which accounts they use, and how often they use them. The discovery phase is live in early access, while blocking and policy enforcement at the browser level are targeted for early access on July 29.

Prompt and tool call analysis inspects prompts, AI responses, and tool calls against Reco Graph context, including agent identity, risk, connections, MCP server access, and user SaaS and AI activity. Reco is targeting general availability for this capability in October.

Smart Remediation shipped in July and turns findings into specific actions, such as reducing permissions, revoking connections, or updating policies. Reco said it will show AI Runtime demos at Black Hat USA 2026 from August 2 to 7.