By Hamza Faisal

Neo, the AI security control platform, has launched Neoverse, a security intelligence engine that researches the fast-growing universe of AI apps and tells security teams what each one can actually do — and whether it is safe to let it run. The launch, announced on Monday, comes with a free, searchable catalog of more than 4 million AI apps, MCP servers, skills and extensions.

The timing reflects a problem that has outgrown the old app-review playbook. Developers now connect MCP servers to coding agents with a single command, employees add AI browser extensions in seconds, and skills can be imported into an agent and run instantly. Many of those add-ons come from marketplaces where anyone can publish, and attackers have noticed: Neo points to TraderTraitor, the threat actor behind the Binance crypto heist, which specializes in building malicious extensions, IDE apps and skills that now function as malware.

Research at the scale of the AI supply chain

Neoverse tackles the volume problem with an agentic swarm — a coordinated team of specialized autonomous agents that research every AI tool, extension and MCP server published to the VS Code, JetBrains, Cursor and Windsurf marketplaces, GitHub, the MCP registry, and the Chrome, Edge and Firefox web stores. Each app is analyzed three ways: static analysis of binaries, scripts, manifests and instruction files, looking for malicious code, hidden instructions, vulnerable components and risky permissions; runtime analysis in a sandbox that captures what the app really does; and documentation analysis that checks behavior against what the app claims, including its privacy policy and whether the tool trains on your data. The engine also folds in CVE and GitHub Security Advisory data.

Every new version of an app is re-analyzed, so a familiar tool that quietly gains AI capabilities — remote system control, say, or the ability to delete files — is caught on its next release rather than coasting on an earlier approval.

From verdict to enforcement

The output is a clear security verdict backed by evidence a team can check: what the app can do, what it can access, and whether it acts with approval or on its own. Inside the Neo platform, that verdict drives enforcement — when Neoverse determines an AI component is malicious, Neo blocks it on every surface where it appears.

"AI software can be installed in a click, and the apps employees already trust can gain new AI capabilities overnight through a routine update," said Nick Warner, Neo's CEO and co-founder. "Security teams have been tasked with vetting apps and AI tools with little more than a star rating to go on. Neoverse gives them real answers."

Free catalog, built-in enforcement

Anyone can sign in at verse.neo.ai and search the catalog to see each app's verdict, key findings, summary and version history. Neo customers get full Neoverse access embedded in the platform, connecting software discovery directly to policy enforcement.

Neo, founded by SentinelOne veterans and backed by Andreessen Horowitz, Bessemer Venture Partners, Craft Ventures and Merlin Ventures, emerged from stealth in July with $100 million in funding. Neoverse is available now.