An LLM agent ran a database-extortion campaign end to end through an unpatched Langflow server
Sysdig says an LLM-driven operator it calls JADEPUFFER ran a database-extortion campaign on its own, entering through unpatched Langflow flaw CVE-2025-3248. It calls the operation the first documented ransomware run end to end by a model.
Reality
- Evidence50
- Adoption
- Not scored
- Hype gap+25
- Incentives70
- Confidence50
Sysdig's threat research team watched one operator work a marimo notebook host for nine hours with hand-written scripts, and the eight-second jump to a bastion host at the end ran on tooling already staged on disk.
Perspective Coverage
3 publishers
- Builder
- Builder 33%
- Operator
- Operator 60%
- Investor
- Investor 7%
Reality
- Evidence68
- Adoption66
- Hype gap+8
- Incentives72
- Confidence70
Sysdig found a credential-theft crew reselling access to a victim's hosted Claude
The Sysdig Threat Research Team says stolen cloud keys were checked against ten hosted model services and fronted to buyers, with the compromised tenant paying for every prompt at invocation prices.
Reality
- Evidence62
- Adoption30
- Hype gap+22
- Incentives72
- Confidence58
A nuclei template hit Langflow's unauthenticated build endpoint 20 hours after disclosure
One HTTP request runs Python on an exposed Langflow server, and Sysdig's honeypots logged exploit attempts inside a day of the March 17 advisory, before any public proof-of-concept existed. The traffic identified itself as nuclei.
Reality
- Evidence58
- Adoption34
- Hype gap+30
- Incentives78
- Confidence52
Sysdig credits Anthropic's Mythos preview with 181 working Firefox exploits
Every number in the account traces to a single vendor blog post that lists no CVE identifiers and no disclosure dates. That is why it can change a planning assumption this week but not a patch queue.
Reality
- Evidence24
- Adoption18
- Hype gap+62
- Incentives82
- Confidence60
Sysdig clocks an AI-assisted intrusion reaching admin in eight minutes
The vendor's threat-research figures arrive without CVE identifiers, dates or sample sizes, worth flagging because the operational consequence is a patch window attackers shut before any patch gets applied.
Reality
- Evidence34
- Adoption45
- Hype gap+42
- Incentives88
- Confidence38
Kubernetes 1.37: the hardening you did not ask for is the part you have to test
Sysdig counts 19 security-relevant changes in the release. The ones that arrive switched on, led by SELinuxMount at stable, are the ones that can break a running cluster.
Reality
- Evidence58
- Adoption30
- Hype gap+10
- Incentives68
- Confidence54