threatdeck
Live 13:45 UTC ·

Detection library

SigmaHunting query: tune before alertingShadow AI and AI-powered attackers

Web proxy: requests to AI model APIs that do not come from the approved AI gateway

Where AI use must go through a sanctioned gateway (for logging, DLP and key management), direct POSTs to model APIs from other sources are shadow AI or leaked-key use. Replace the example source range in the filter with your gateway's egress addresses.

Log source
Web proxy logs (Sigma proxy category)
Techniques
LLM02 Sensitive Information Disclosure · AML.T0096 AI Service API
MITRE ATT&CK
T1567Exfiltration Over Web Service
Reviewed
2026-11-10

Query

title: AI Model API Request Not From Approved Gateway
id: 634cab20-7a79-41db-a7f3-f55675fb6f1f
status: experimental
description: Detects POST requests to public AI model APIs from sources other than the approved AI gateway.
references:
    - https://platform.claude.com/docs/en/api/overview
    - https://developers.openai.com/api/reference/overview
    - https://aistudio.google.com/api-reference
    - https://learn.microsoft.com/en-us/azure/ai-foundry/openai/reference
author: ThreatDeck
date: 2026-11-09
modified: 2026-11-10
tags:
    - attack.exfiltration
    - attack.t1567
    - detection.threat-hunting
logsource:
    category: proxy
detection:
    selection_method:
        cs-method: 'POST'
    selection_host:
        - cs-host:
              - 'api.openai.com'
              - 'api.anthropic.com'
              - 'generativelanguage.googleapis.com'
              - 'api.deepseek.com'
              - 'openrouter.ai'
              - 'router.huggingface.co'
        - cs-host|endswith: '.openai.azure.com'
    filter_main_gateway:
        src_ip|cidr: '192.0.2.0/24'  # replace with the AI gateway egress range
    condition: all of selection_* and not filter_main_gateway
falsepositives:
    - Approved applications that call providers directly
level: low

Expected false positives

  • Approved applications that call providers directly instead of via the gateway

Sources

Reviewed against vendor documentation, not tested in your environment. Defensive use only.

Following

Follow the vendors, products, AI tools, techniques and actors you care about. Matching stories, CVEs and exploited bugs are marked, a Following tab appears in the feed, and returning visits show what changed. Saved only in this browser.

Presets

Get updates

A personal RSS feed of new stories on everything you follow. Paste it into any feed reader, Slack or Teams RSS app.