Elmoz maps where AI agents and machine identities can reach.

#Why Elmoz exists

Companies now run more machine identities than people. AI agents add to that fast, and they chain tools and credentials on their own. Most security programs can list these identities. Very few can say where each one leads. The incidents we analyse on the research blog keep showing the same shape: a token or service account that looked harmless on its own, and a short chain of legitimate steps from it to production data.

Elmoz answers one question for security and platform teams: from this agent or identity, what can an attacker eventually reach, and where is the cheapest place to break the path?

#What we build

The platform overview explains each part in detail.

1. Discovery. An inventory of AI agents and non-human identities across cloud, SaaS and developer tooling, including agents nobody registered.
2. Access mapping. What each identity can do: roles, scopes, tokens and inherited rights.
3. Attack paths. Which chains of access lead from an agent or identity to sensitive data, and the breakpoint that removes the most risk.
4. Evidence. An audit trail of agent activity and of the decisions your team makes about them.

#Founder

Sascha Buhle founded Elmoz in 2025 and writes the incident teardowns on this site. He holds a B.Sc. in IT Security from TH Brandenburg and has worked as a security consultant on information security management and regulation, including ISO 27001, NIS2 and the EU AI Act. He is a certified Professional Scrum Master (PSM II) and Professional Agile Leader (PAL I).

Connect on LinkedIn or write to [email protected].

#Company facts

Want to see where your agents can reach? Book a demo.