SecurityBrief India - Technology news for CISOs & cybersecurity decision-makers
India
Orca adds tools to spot AI-built apps in the cloud

Orca adds tools to spot AI-built apps in the cloud

Wed, 18th Nov 2026 (Today)
Shadda Lie Catugal
SHADDA LIE CATUGAL Editor

Orca Security has added tools to its platform to identify AI-built applications and monitor autonomous AI agents in cloud environments. The update targets software created outside standard security processes.

The changes are designed to help security teams find applications built with consumer and developer AI tools that may already be running without formal review. They also aim to show what data those applications can access and whether AI agents have broader permissions than necessary.

Employees across business functions are increasingly using services such as Lovable, Replit and Bolt to create software without going through conventional engineering teams. At the same time, developers are relying more heavily on autonomous coding agents to write code and carry out tasks, creating new challenges for cloud security teams.

Orca said its new app discovery features look for digital traces left by AI application builders across DNS records, workloads, code repositories, cloud identities and other parts of the environment. Those signals include DNS records linked to builder infrastructure, installed packages, file markers, commit identities and cloud permissions.

The system is intended to detect applications built with tools including Lovable, v0, Replit, Base44 and Bolt, as well as associated hosting and data services such as Vercel, Netlify and Supabase. Findings are presented through a service fingerprints view, asset-level details and an AI app generation dashboard.

Security teams often know which approved platforms they monitor, but not which services employees may have used without registration or disclosure. Orca's approach focuses on finding evidence of activity already present in the environment, rather than relying on developers or business users to declare what they have built.

"AI has turned everyone into a builder, but security teams are often the last to know what's been created," said Gil Geron, Chief Executive Officer and Co-Founder of Orca Security.

"An employee can deploy an application over the weekend, while an autonomous agent can access sensitive data and take actions across cloud environments. Organisations need to understand not just what exists, but what it can access and where it creates risk. We're bringing that context to the entire AI development ecosystem so companies can keep innovating without losing control," Geron said.

Agent oversight

Alongside app discovery, Orca has expanded its monitoring of AI agents, including cloud-managed and self-hosted tools. This includes coding agents such as Claude Code, Codex and Gemini CLI running on cloud workloads.

A new agents tab within Orca's AI Bill of Materials is intended to provide a central inventory of agents alongside AI models, packages and MCP servers. Users can filter that inventory by cloud, account and framework.

The platform also maps links between agents, the models they use, the tools and servers they connect to, the cloud identities under which they operate and the data stores they can access. This is intended to help security teams judge how far an agent's reach extends and where sensitive data might be exposed.

Risk checks in the updated platform include identifying agents with excessive permissions, disabled approval controls, exposed credentials in MCP configurations, insecure HTTP connections and access to destructive tools. Those findings are folded into Orca's risk scoring and attack path analysis.

The release also covers AI-generated code. Organisations will be able to attribute code to the coding agents and AI models that produced it, giving security teams a view of how much AI-generated code exists across repositories and which tools are contributing the most risk.

Growing concern

The changes reflect a broader shift in software creation as AI tools move beyond engineering teams and into wider business use. That trend has increased the number of people able to create and deploy applications quickly, while making it harder for companies to track what has been put into production.

For security teams, the issue is not only that an unapproved application exists, but also whether it can reach internal systems, cloud resources or sensitive corporate data. Autonomous agents create a related problem because they can carry out actions using the permissions and credentials they inherit.

Grace Trinidad, Research Director at IDC, said discovery remains a central challenge for many organisations.

"Many organisations still struggle to discover what's been built and where it runs. Now that building is no longer confined to engineering, discovery has to start from what these AI-generated apps leave behind in the environment, not from the builder declaring anything," Trinidad said.

Orca is also extending support in its AI application security tools to Base44, OpenAI and Vercel, while adding posture assessments for issues such as disabled row-level security in Supabase projects. The broader aim is to bring AI-built applications and autonomous agents into the same cloud security workflows already used for vulnerabilities, misconfigurations, identities and sensitive data.