The transition of artificial intelligence from passive recommendation engines to active operational agents is forcing a fundamental shift in enterprise IT governance and security requirements. ScienceLogic is addressing this shift by introducing Skylar AI 2.7, a release focused on providing IT administrators with the visibility and granular controls necessary to manage autonomous agentic functions. Announced at Nexus Live 2026, the update targets the growing need for oversight as AI agents gain more authority within complex IT environments. By implementing new safeguards for knowledge ingestion and tool selection, the company aims to provide a framework where agentic AI can scale without compromising organizational security or operational accountability.
New Security and Governance Controls in Skylar AI 2.7
ScienceLogic is positioning Skylar AI 2.7 as a mechanism for managing the risks associated with autonomous AI agents by introducing specific oversight features for agent routing and tool selection. The update includes a new agent-routing trace and a tool catalog, which allow administrators to observe how the Skylar Advisor routes specific requests and which internal or external Model Context Protocol (MCP) tools are considered during the process. This level of transparency is intended to help IT teams understand the decision-making logic of their AI agents. Furthermore, the release introduces the ability for administrators to manage agentic functions and MCP tools independently. This means specific functions or tools can be disabled without disrupting other approved agents or existing workflows, providing a level of surgical control over automated processes.
To protect the integrity of the data informing these agents, the company has introduced an "upload gatekeeper." This feature inspects documents for potentially risky content—such as exposed secrets, malicious instructions, or hidden-text manipulation—before they are integrated into the AI's knowledge base. If suspicious content is detected, the system can quarantine the documents for review by a subject-matter expert. This is supported by human-in-the-loop review workflows designed to govern how content is evaluated and approved. Additionally, Skylar Advisor now allows users to suppress specific advisory patterns, enabling teams to refine the relevance of AI-generated recommendations based on historical user actions.
Strengthening the Skylar Advisor Knowledge Foundation
The Skylar AI 2.7 release builds upon a rapid development cycle for ScienceLogic’s AI capabilities, following the introduction of Skylar AI 2.6 in September and Skylar AI 2.5 in August. The current iteration focuses heavily on the "knowledge foundation" that drives Skylar Advisor. As AI agents move toward taking direct action in IT operations, ScienceLogic is emphasizing the need for a secure and governed data environment. The company is providing tools that allow organizations to set clear boundaries around agent activity, ensuring that the information used to generate insights is both accurate and secure.
According to ScienceLogic CEO and cofounder Dave Link, the market is moving from AI that merely provides recommendations to AI that takes action. This evolution requires IT leaders to have the ability to audit how agents operate and understand the specific information driving those operations. By providing visibility into agentic functions and the ability to implement what Chief Product Officer Michael Nappi describes as a "smart agentic kill switch," ScienceLogic is attempting to build a foundation for scaling agentic AI. These capabilities are designed to help organizations maintain governance and trust even as they grant AI agents more autonomy within their digital infrastructure and enterprise IT ecosystems.
Key Takeaways
- Skylar AI 2.7 introduces an upload gatekeeper to inspect documents for malicious instructions, exposed secrets, and hidden-text manipulation before they enter the AI knowledge base.
- Administrators can now manage agentic functions and Model Context Protocol (MCP) tools independently, allowing for the disabling of specific tools without interrupting other workflows.
- The update provides new visibility through an agent-routing trace and a tool catalog to show how Skylar Advisor routes requests and selects internal or external MCP tools.
TechInsyte's Take
In our view, ScienceLogic’s focus on "agentic controls" signals a critical realization in the enterprise AI market: the primary barrier to AI adoption is no longer capability, but controllability. As AI moves from a consultative role to an operational one, the risk profile shifts from "incorrect advice" to "unauthorized action." By prioritizing features like the upload gatekeeper and granular MCP tool management, ScienceLogic is attempting to solve the "black box" problem that prevents CIOs from fully delegating tasks to autonomous agents. The introduction of a "kill switch" concept and human-in-the-loop workflows suggests that the industry is moving toward a hybrid model where AI autonomy is strictly bounded by human-defined governance. This approach is a necessary step for any organization looking to move beyond pilot programs into production-grade agentic AI.
Questions & Answers
How does Skylar AI 2.7 prevent malicious data from influencing AI-generated guidance?
The platform utilizes a new "upload gatekeeper" that inspects documents for risky content, including hidden-text manipulation, malicious instructions, and exposed secrets. Suspicious documents are quarantined for subject-matter expert review to ensure only governed content informs the AI.
Can administrators disable specific AI functions without stopping all automated workflows?
Yes. Skylar AI 2.7 provides granular control that allows administrators to manage agentic functions and Model Context Protocol (MCP) tools independently. This enables the disabling of specific functions or tools while allowing other approved agents and workflows to continue operating.
What visibility does the update provide regarding how AI agents make decisions?
The release introduces an agent-routing trace and a tool catalog. These features allow administrators to see exactly how Skylar Advisor routes a specific request and which internal or external MCP tools the agent considers during its process.
What is the strategic purpose of the new advisory suppression feature in Skylar Advisor?
The feature allows users to suppress advisory patterns that have been determined to not require further attention. This is intended to help IT teams focus on more relevant recommendations by filtering out noise based on user action.
Source: ScienceLogic
