Jemimus / Wikimedia Commons / CC BY 2.0
AUSTIN, Texas – Orca Security announced new platform capabilities Tuesday that it says can uncover AI-built applications security teams do not know about, inventory autonomous agents and assess their access to company data.
The November 10 announcement expands Orca’s AI Security offering across cloud environments and code repositories. The capabilities are company descriptions, not independently verified evidence that every AI application or security risk will be found or prevented.
So-called shadow AI applications are built outside the processes a security team normally monitors. Orca says its AI AppGen Security capabilities look for digital fingerprints left by builders such as Lovable, v0, Replit, Base44 and Bolt, as well as hosting and data services including Vercel, Netlify and Supabase.
Those signals can include DNS records, installed packages, file markers, builder commit identities and cloud permissions. Orca says it connects them with cloud identities and accessible resources so security staff can judge which applications may reach sensitive data. Its agentless SideScanning technology is designed to find previously unknown services without requiring employees to register them.
“AI has turned everyone into a builder, but security teams are often the last to know what’s been created,” CEO and co-founder Gil Geron said in the release.
For autonomous agents, Orca says a new Agents tab in its AI Bill of Materials inventories cloud-managed and self-hosted agents, including coding agents operating on cloud workloads. It maps their relationships with models, tools, servers, cloud identities and data stores.
The company says the platform can flag excessive permissions, disabled approval controls, exposed credentials, insecure connections and access to destructive tools. Findings feed into its existing risk scoring and attack-path analysis to help security teams decide which exposures to address first.
Orca also says it can attribute AI-generated code to the agents and models that produced it. Its AI Skills Hub is designed to bring security requirements, scanning and suggested fixes into coding-agent workflows before code is committed.
What it means: The announcement targets a visibility problem: staff and AI agents can create software or gain access before security teams know they exist. For businesses, the test is whether discovery and permissions checks work in their own environments. The release does not provide pricing or measured customer outcomes.
JBizNews Desk | Technology & Israel Business
© JBizNews.com. All rights reserved. Unauthorized reproduction or redistribution is strictly prohibited.
