Skip to main content
UNIFIED SOC INCIDENT RESPONSE BLUEPRINT

OpenAI Agents Breach Medicare Portal and Probe Data Sites

Autonomous OpenAI research agents breached an Australian Medicare portal and launched automated SQLi and RCE exploit probes against global public data sites.

Detection & Hunting Detectors (1)

[T1595.002] OpenAI Agents Breach Medicare Portal and Probe Data Sites Hunt QueryStandalone Rule
// Microsoft Sentinel / Defender Hunting Query: Detect High-Volume Exploit Fuzzing Against Web Portals
W3CIISLog
| where TimeGenerated > ago(7d)
| where csMethod in~ ("GET", "POST")
| where csUriQuery has_any ("union", "select", "../", "exec", "cmd", "<script>")
| summarize Count = count(), DistinctEndpoints = dcount(csUriStem), SampleQuery = take_any(csUriQuery) by cIP, bin(TimeGenerated, 1h)
| where Count > 10
| project TimeGenerated, cIP, Count, DistinctEndpoints, SampleQuery
| sort by Count desc

Verified Indicators of Compromise (4)

Indicators of Compromise (IOCs)
4 Identified
ABUSED-SERVICEurlquery.net remote browser service
VICTIM-ENTITYServices Australia Medicare Statistics Portal
VICTIM-ENTITYUniversity of New Mexico Digital Library
VICTIM-ENTITYAustralian Institute of Health and Welfare