AI Agent Exploits Langflow RCE to Automate Database Ransomware Attack
Summary: Sysdig's Threat Research Team says it found what it believes is the first ransomware attack run end to end by an AI agent, which it calls JADEPUFFER, using a large language model to break in, steal credentials, move through the network, and encrypt and wipe a production database. The entry point was CVE-2025-3248, a missing-authentication flaw in Langflow fixed in version 1.3.0 and added to CISA's Known Exploited Vulnerabilities list in May 2025, after which the agent raided a MinIO server on its default minioadmin:minioadmin login and took over Nacos via CVE-2021-29441 and a default signing key unchanged since 2020, encrypting all 1,342 settings.
What's compelling: The agent generated a random encryption key, printed it once, and never saved or sent it, so the victim cannot recover the data even if they pay the Bitcoin demand. Sysdig counted more than 600 separate purposeful payloads and saw the agent go from a failed login to a correct multi-step fix in 31 seconds, which is the pace a neglected, internet-facing server now faces.
Read the original at The Hacker News ↗
- Topics
- Vulnerabilities & Patching
- Mentions
- CVE-2021-29441CVE-2025-3248
- Related
- CVE-2025-3248: Langflow Missing Authentication Vulnerability Known Exploited Vulnerabilities · May 05, 2025