← Glossary Company
Oasis Security
A security research firm focused on non-human identities, machine accounts and the credentials that AI agents use.
Oasis Security works on a corner of security that got much larger with AI: non-human identity. That means the accounts, tokens and API keys used by software rather than people, which now includes every AI agent doing work on someone’s behalf. Those credentials are typically created quickly, shared widely and rarely rotated, which makes them an attractive target.
The company’s research team publishes findings on AI tooling, and in August 2026 it disclosed CVE-2026-65105, the NemoClaw flaw that let a single malicious web page plant permanent hidden instructions inside a locally running model.
Mentioned in