Netskope Inc. (NASDAQ:NTSK) has announced Skylight Agent Action Control, a capability that helps security teams dictate AI agent behavior by classifying their risky actions and controlling them before execution. The latest tool caters to a pressing industry gap, as 91% of enterprises today lack capabilities to timely mitigate such risky attempts by agents. Moreover, 54% of the organizations using AI agents have witnessed a suspected or confirmed security breach during the past 12 months.
Skylight Adds Real-Time Agent Action Controls
Skylight Agent Action Control backs Netskope’s push to become a leader across the AI agent governance segment. The tool offers a proactive approach towards risk management rather than a reactive response. The development takes place at a time when Netskope’s AI security portfolio is undergoing rebranding from Netskope One AI Security to Netskope Skylight, which enhances the overall projection of the company’s comprehensive AI security offerings. The portfolio now includes AI Guardrails, AI Command Center, Agentic Broker, AI Gateway, and AI Red Teaming.
The announcement follows management’s ongoing momentum of introducing innovative features to its existing products, along with new offerings. During the recently reported second quarter, it came up with Netskope One DataSec Command Center, which is a unified platform for tracking and protecting sensitive data across endpoints, cloud, AI environments, and more. The company also introduced some new features to its NewEdge AI Fast Path capability, which was shown to reduce latency by as much as 90% to popular AI destinations, as per company testing.
To understand the importance of proactive agent governance, it is important to highlight the risks Skylight Agent Action Control aims to address. The first is ‘authority drift,’ where an AI agent’s permission footprint expands beyond what was originally approved. Separately, vague or overly permissive instructions can lead agents to attempt high-risk actions, such as deleting a production repository.
There is also a widespread inability to intervene in real time. As highlighted above, a large number of organizations are unable to avoid a risky agent action in a timely manner, which leaves security teams reactive rather than proactive. Such risks indicate a critical gap between enterprise oversight and agent autonomy, especially when it comes to actions like unauthorized external communication and data destruction.
However, the investment payoff will depend on how quickly enterprises adopt Agent Action Control and whether Netskope can monetize the broader Skylight portfolio in an increasingly competitive AI security market.
Institutional Sentiment
Based on data tracked across 1,000+ hedge funds by Insider Monkey, number of smart-money managers with long-term exposure in the stock has been going up. A total of 35 hedge funds held positions by the end of Q2 2026, compared to 31 in Q1 2026. Short interest is hovering below 5%, which shows limited bearish skepticism.
ICONIQ Capital is the largest institutional stakeholder, as per Yahoo Finance database, holding 66.27 million shares as of June 30. This amounts to 24.81% of the outstanding shares. Other notable institutional investors include BlackRock and Canada Pension Plan Investment Board, holding 3.48% and 3.35% of outstanding shares respectively.
What Lies Ahead
Since Agent Action Control operates on existing network traffic already inspected by the Netskope platform, customers gain this new layer of security without deploying additional tools or consoles. This strengthens Netskope’s value proposition by enhancing platform stickiness and could deepen its competitive edge in AI security as enterprises look to address growing concerns around agent authority drift.
READ NEXT: 12 Best Industrial Stocks With More Than 50% Upside and 10 Best Stocks Under $10 That Could Triple.
