Analysis
What Is Agentic worm? Definition & Examples
Publication date unknown · Discovered nhimg.org
Page published
Publication date unknown · First observed: 12 Oct 2026
Coverage timeline
Single-source analysis — one report is available.
Why it matters
Agentic worms exploit default trust between autonomous agents, so defenders building multi-agent or A2A systems need to treat every handoff as a policy decision to limit blast radius.
A glossary entry from NHI Management Group defines the "agentic worm" as a hypothetical malware pattern in which one AI agent induces another to run malicious or unsafe prompts, spreading through inter-agent communication, delegated tool use, and shared memory rather than classic self-replication. The reference piece explains propagation paths, why trust boundaries between agents fail, and where containment and governance matter most.