AI agent went from marimo notebook exploit to full database dump in under an hour

Undisclosed, disclosed May 2026. AI-run attack Data exposed High

Sysdig observed an LLM-driven intrusion that exploited an internet-facing marimo notebook, harvested cloud credentials, pulled an SSH key from AWS Secrets Manager and exfiltrated a full PostgreSQL database through four pivots in under an hour, without pre-written scripts.

Disclosed

May 2026

Organization

Undisclosed

Vendor

Unknown

Kind of AI

Autonomous agent

AI's role

Attacker used AI

How it happened

AI-run attack: Autonomous intrusion

Harm

Data exposed

Data involved

Credentials, Unknown

Reach

One organization

Severity

High

Model at fault

No

Status

Confirmed

Vulnerability IDs

CVE-2026-39987

Sources

1. https://www.sysdig.com/blog/ai-agent-at-the-wheel-how-an-attacker-used-llms-to-move-from-a-cve-to-an-internal-database-in-4-pivots