Observed incident

PoeLLM botnet compromises exposed AI services

Reported

What happened

Black Lotus Labs traced a growing cryptomining botnet compromising exposed services including LiteLLM and Ollama, with command-and-control locations concealed in a changing online poem.

Parties and reporting

AI tools and systems: PoeLLM, LiteLLM, Ollama, XMRig, Iron

Outcome

Lumen blocked known C2 infrastructure. Its researcher subsequently reported disruption after the poem repository became unavailable; complete removal from compromised hosts is unconfirmed.

Operating lesson

Patch and authenticate internet-facing AI services, and detect mining processes and unusual outbound connections.

Evidence and limits

The campaign dates to April. AI infrastructure was targeted; the report does not establish that an LLM operated the malware. The later Lumen statement appears in an undated update to the October 7 report; it is not proof of eradication from every infected host.

Sources