Celigo Launches MCP Management Platform for Enterprise AI Agent Governance

Intelligent automation startup Celigo Inc. announced its Model Context Protocol Management platform today, providing a new mechanism for information technology and security teams to govern artificial intelligence assistants. The platform is designed to control how these autonomous agents interact with and act upon enterprise systems, addressing emerging governance gaps.

The announcement targets the growing use of the Model Context Protocol, or MCP, which has become a standard for connecting AI assistants to business services. According to SiliconANGLE, many enterprise services now host MCP servers that expose tools allowing agents to interact with both basic and complex controls within an organization’s infrastructure.

Context on Model Context Protocol

Anthropic PBC introduced the Model Context Protocol in 2024 as an open-source standard, often described as a USB-C port for AI. This protocol acts as a universal access point, allowing AI models and agents to plug into databases, local files, search engines, and application programming interfaces to access necessary resources.

Examples of agent activities facilitated by MCP include looking up customer information in management systems, checking invoices in enterprise resource planning systems, updating order statuses in parcel tracking systems, or working through tickets on service desks. Tasks that previously required integration specialists can now be completed autonomously by AI agents such as Claude, ChatGPT, Copilot, and Cursor.

Security Risks and Governance

The autonomy of these agents introduces potential security risks and access issues if they possess permissions they should not have or take actions outside their designated roles. Celigo noted that it is important to carefully limit agent activity and guide their operational boundaries, as AI agents can hallucinate or make mistakes with little supervision.

Additionally, these agents can be hijacked by malicious actors if their access is not properly secured. Chief Product Officer Matt Graney stated that AI assistants represent a new class of enterprise actors, noting that the old model of shared credentials and broad access was never designed for them.

Who Is Affected

Enterprises deploying AI agents that interact with internal business systems are directly affected by these governance challenges. Organizations need to know who is behind every AI request, what it is authorized to do, and what actions it has taken, according to Graney. This applies to any IT or security team managing access to databases, APIs, or business applications via AI tools.

Celigo’s MCP Management platform provides a single governed path to lock down agents and keep them within their respective domains. The platform ensures that every call traces back to a real person and that access

What to do and how to stay safe

• Sudden role changes in output signal a successful override of system instructions.
• Hidden metadata in images or documents often carries invisible injection payloads.
• Immediate isolation of the affected session prevents lateral movement within the AI workflow.

Prompt injections exploit the model’s ability to follow instructions by hiding malicious commands in input data. Isolate affected sessions immediately and implement external guardrails to filter inputs before they reach the model.

Step-by-step guide: Prompt Injection Warning Signs: How to Spot AI Manipulation

General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.