{
  "id": 8243636,
  "title": "WSO2 Releases Agent Manager as Enterprises Look to Control Growing AI Agent Sprawl",
  "url": "https://urgent.news/2026/09/18/wso2-releases-agent-manager-as-enterprises-look-to-control-growing-ai",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-18T12:00:00.000Z",
  "source": {
    "name": "InfoQ",
    "slug": "infoq",
    "url": "https://www.infoq.com/news/2026/09/ws02-agent-manager/"
  },
  "original_language": "en",
  "account": "WSO2 has made the general availability of its Agent Manager platform official, a solution aimed at helping enterprises manage the surge in AI agents across various models and environments. This open-source tool addresses the fragmented management of agent governance, identity, security, and operations that organizations face.\n\nAgent Manager separates the management of agents from their logic, allowing for consistent controls regardless of the underlying model, framework, or runtime. It enters general availability after a successful beta period in June 2026, bringing enhancements like more robust agent identity features, MCP interaction controls, and a Kubernetes-native, sandboxed execution environment.\n\nCentral to the platform is the management of agent identity and authorization, which is crucial given the complex interactions agents can have, such as invoking tools, accessing APIs, delegating tasks, and interacting with other agents. It offers a centralized inventory system with features like verified agent identities, role-based access, delegation capabilities, token exchange, and the ability to revoke access.\n\nThe platform also includes more than 40 pre-built controls spanning areas like masking personally identifiable information and rate limiting, which can be applied across different parts of an agent workflow. A sandboxed runtime, built on Kubernetes, provides a safe space for agent execution, while OpenTelemetry is used for tracing and evaluating agent behavior over time. Agent Manager supports a wide range of technologies, including LangChain, CrewAI, Amazon Bedrock, Azure, Ballerina, and custom agents, with a focus on maintaining independence from specific AI models, frameworks, or providers. This flexibility is key in an industry where AI technologies are rapidly evolving.\n\nThe release reflects a broader industry trend where governance is becoming a platform concern rather than just an AI development issue. Major cloud providers like AWS and Microsoft are also adding layers of non-human identity, tool authorization, policy enforcement, runtime isolation, and observability to their agent platforms. Given the diversity of approaches, the development of a framework-independent control plane for managing agents across multiple models and cloud platforms is increasingly seen as necessary for organizations operating at scale.",
  "summary": "WSO2 has announced the general availability of WSO2 Agent Manager, an open-source platform designed to provide centralized governance, identity management, security controls, and operational oversight for AI agents running across different models, frameworks, and deployment environments. By Craig Risi",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}