Observed incident

JADEPUFFER chains AI-assisted intrusion and destructive extortion

Reported

What happened

Sysdig documented an operation exploiting Langflow, harvesting credentials and attacking a downstream database, with runtime behavior that researchers assessed as autonomous AI-driven extortion.

Parties and reporting

AI tools and systems: Unspecified LLM agent, Langflow, MySQL, Alibaba Nacos, ENCFORGE, Microsoft Azure

Outcome

Sysdig observed destructive database operations; subsequent reports linked the actor to ransomware targeting AI artifacts and deletion of Azure resources.

Operating lesson

Patch AI frameworks, restrict workload identities and protect backups with controls separate from production administration.

Evidence and limits

AI autonomy is Sysdig's assessment from runtime behavior; the model was not identified. Data exfiltration was not independently verified.

Sources