Agent Marketplace News

AI agent marketplaces are becoming the storefronts and distribution layer for autonomous software. Instead of building every capability from scratch, organizations can browse catalogs of prebuilt agents, tools, and connectors, then plug them into their own workflows. Many of these listings rely on the Model Context Protocol (MCP), an open standard that lets AI models call external tools and data sources in a consistent way.

Why it matters now

The category is moving quickly from experiment to enterprise purchase. Security vendors, AI labs, and open-source projects are launching or expanding marketplaces, and pricing is evolving as shared credit pools and consumption-based billing appear alongside traditional licenses. At the same time, the connective tissue underneath is under heavy scrutiny. Researchers keep surfacing weaknesses in how MCP servers handle local execution, authentication, and authorization, and debate continues over whether some risks are deliberate design choices or genuine flaws. As agents gain the ability to act on real systems, a single vulnerable connector listed in a popular marketplace can expose many customers at once.

What you'll find here

This hub tracks:

• Marketplace launches and business models, including how vendors package, price, and meter agent capacity
• Protocol and spec developments, such as MCP draft specifications and which platforms support them
• Security research and vendor responses covering code execution risks, OAuth implementation issues, and SDK-level exposure
• Developer tooling from major AI platforms, including agent-building kits and guardrail features
• Funding and enterprise strategy among companies betting on agent deployment
• Open-source frameworks and how they compete with or feed into commercial catalogs

Whether you evaluate agents for your organization or build them, check back for coverage of how this ecosystem is taking shape and where it still breaks.