Tia N.List

CTI · reviewed · 2026-08-28

Hackers Target LiteLLM, RAGFlow and Kestra AI Infrastructure to Steal API Keys and Mine Crypto

Full story & timeline

Actors

Unknown

Malware / tools

—

Campaigns

AI infrastructure targeting

CVEs

CVE-2026-42271CVE-2026-48710

Victim sectors

AItechnology

Geography

—

MITRE ATT&CK

Indicators

No curated indicators tied to this case.

Detections

Detection content derived from this case analysis. Status: experimental — not validated against live telemetry.

Sigma raw

title: AI Infrastructure Compromise - Environment Variable Exfiltration
id: e7a1b4c6-9c5f-4e0d-cf8a-4f6c1e3d5a7b
status: experimental
description: Detects potential compromise of AI gateway infrastructure (LiteLLM, RAGFlow, Kestra) through environment variable harvesting from /proc/1/environ and anomalous database access patterns. Based on Microsoft and GBHackers reporting on AI infrastructure targeting campaigns.
references:
  - https://cyberpress.org/hackers-target-litellm-ragflow-kestra-ai
  - https://www.microsoft.com/en-us/security/blog/ai-infrastructure-targeting
author: Tia N. List
date: 2026-08-28
tags:
  - attack.t1190
  - attack.t1552
  - attack.t1041
logsource:
  category: process_creation
  product: linux
detection:
  selection_environ_access:
    EventID: 1
    CommandLine|contains:
      - '/proc/1/environ'
      - 'cat /proc/1/environ'
      - 'strings /proc/1/environ'
  selection_env_harvest:
    EventID: 1
    CommandLine|contains:
      - 'OPENAI_API_KEY'
      - 'ANTHROPIC_API_KEY'
      - 'LITELLM_MASTER_KEY'
      - 'AZURE_API_KEY'
      - 'GEMINI_API_KEY'
      - 'DATABASE_URL'
  selection_db_access:
    EventID: 1
    CommandLine|contains:
      - 'psql'
      - 'pg_dump'
      - 'SELECT * FROM model_config'
      - 'SELECT * FROM virtual_key'
      - 'litelimit.postgresql'
  condition: selection_environ_access or selection_env_harvest or selection_db_access
falsepositives:
  - Legitimate AI platform administration
  - Authorized database maintenance operations
level: critical

References

Source on GitHub