Loading threat actor…
JADEPUFFER is an emerging, unattributed ransomware and extortion threat actor assessed as the first publicly documented agentic threat actor to direct most of an end-to-end ransomware intrusion through large-language-model-driven automation. The actor is financially motivated and has repeatedly targeted internet-exposed Langflow deployments as an entry point, primarily by exploiting CVE-2025-3248 to obtain unauthenticated code execution. From that foothold, JADEPUFFER has been observed conducting reconnaissance, harvesting credentials and secrets, probing internal services, establishing persistence, escalating privileges, and moving laterally into production environments. Documented JADEPUFFER activity includes compromise of Langflow-associated infrastructure, dumping backing databases, accessing object storage with weak or default credentials, pivoting into MySQL and Alibaba Nacos environments, and carrying out destructive extortion actions. In an earlier campaign, the actor encrypted and destroyed database and configuration data, including Nacos records, and created ransom artifacts, with reporting indicating the encryption workflow may not have preserved recoverable keys. Subsequent activity showed operational evolution from ad hoc scripts and native database encryption functions to a purpose-built Go ransomware payload named ENCFORGE. ENCFORGE is designed to encrypt AI and machine-learning assets, including model checkpoints, training datasets, vector indexes, embeddings, and related artifacts. In observed attacks, JADEPUFFER used exposed Docker socket access to create a privileged container escape path and deploy ENCFORGE onto the host after adapting to payload-delivery failures. The ransomware uses hybrid cryptography and selectively targets a broad set of AI-specific file types. Reporting found no verified evidence of data exfiltration or a leak site in the ENCFORGE campaigns, but JADEPUFFER has been associated more broadly with encryption, data theft, and extortion targeting exposed AI platforms. A defining characteristic of JADEPUFFER is rapid, machine-speed adaptation during intrusion operations. Researchers observed self-narrating payloads, iterative script generation, autonomous retry logic, and quick correction of failed steps, including authentication and deployment errors. Public reporting does not link JADEPUFFER with high confidence to any known ransomware-as-a-service program, established criminal cluster, or nation-state actor. The actor is best understood as a financially motivated, AI-enabled ransomware operator focused on AI-adjacent infrastructure and high-value enterprise AI assets.
Mallory correlates actor tradecraft and target patterns against your stack, your sector, and your geography. See overlap before they land.
Who, where, and (when attributed) which flag flies behind the operation. Pulled from open-source reporting and Mallory's analyst review.
Sectors the actor has been observed targeting.
Attributed origin per open-source reporting.
37 distinct techniques observed across reporting, grouped by tactic. Hover any cell for the evidence excerpt; click through for MITRE's full description.
2 malware families attributed to this actor across reporting.
2 CVEs this actor has used in observed campaigns. 2 of them exploited in the wild.
To protect themselves from JADEPUFFER and ENCFORGE attacks, Sysdig recommended organizations patch the Langflow vulnerability tracked as CVE-2025-3248.
Exploitation de CVE-2021-29441 (bypass auth Nacos) et forge de JWT via la clé de signature par défaut de Nacos.
11 indicators attributed to this actor: domains, IPs, hashes, and other artifacts pulled from reporting. View more in app.
20 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.
Targeted exposed AI and product-lifecycle platforms for encryption, data theft, and extortion.
A ransomware operation referenced as using an AI agent to automate the full intrusion lifecycle, including reconnaissance, credential theft, lateral movement, privilege escalation, and data encryption.
Agentic ransomware operations targeting AI assets, including trained models, datasets, and related artifacts, by exploiting Langflow and abusing exposed Docker sockets to deploy ENCFORGE for encryption.
An agentic threat actor conducting end-to-end extortion operations against AI infrastructure and evolving to use ransomware specifically to destroy AI models and related enterprise AI assets.
Match sector + geo + tech-stack targeting against your real footprint.
Every observed MITRE ATT&CK technique, grouped by tactic.
Families this actor is known to deploy, with IOCs and behavior.
CVEs this actor has used in known campaigns.
YARA, Sigma, Snort, and vendor rules, auto-deployed to your SIEM.
Domains, IPs, and hashes tied to this actor, refreshed continuously.