threatdeck
Live 13:45 UTC ·

Detection library

CloudTrail Lake SQLHunting query: tune before alertingCloud AI abuse: Amazon Bedrock (LLMjacking)

Bedrock reconnaissance burst: model, availability and logging checks in one hour

Before using a stolen key, LLMjacking tooling enumerates models (ListFoundationModels, ListInferenceProfiles), checks availability and agreement offers, and reads the logging configuration (GetModelInvocationLoggingConfiguration). This query finds principals that call several of these APIs, including at least one availability or logging check, within the same hour.

Log source
AWS CloudTrail Lake event data store (management events)
Techniques
LLM10 Unbounded Consumption · AML.T0034 Cost Harvesting · AML.T0040 AI Model Inference API Access · AML.T0069 Discover LLM System Information
MITRE ATT&CK
T1526Cloud Service DiscoveryT1078.004Cloud Accounts
Reviewed
2026-11-10

Query · CloudTrail Lake SQL

SELECT userIdentity.arn AS principal, sourceIPAddress,
       date_trunc('hour', eventTime) AS hour_bucket,
       count(DISTINCT eventName) AS distinct_apis,
       array_agg(DISTINCT eventName) AS apis,
       count(*) AS calls
FROM $EDS_ID
WHERE eventSource = 'bedrock.amazonaws.com'
  AND eventType = 'AwsApiCall'
  AND eventName IN ('ListFoundationModels', 'GetFoundationModelAvailability', 'ListFoundationModelAgreementOffers',
                    'ListInferenceProfiles', 'GetModelInvocationLoggingConfiguration', 'GetUseCaseForModelAccess',
                    'InvokeModel', 'InvokeModelWithResponseStream', 'Converse', 'ConverseStream')
  AND eventTime > date_add('day', -7, now())
GROUP BY userIdentity.arn, sourceIPAddress, date_trunc('hour', eventTime)
HAVING count(DISTINCT eventName) >= 3
   AND count_if(eventName IN ('GetModelInvocationLoggingConfiguration', 'GetFoundationModelAvailability',
                              'ListFoundationModelAgreementOffers')) > 0
ORDER BY distinct_apis DESC, calls DESC

Other platforms

Microsoft Sentinel / Defender (KQL)

Microsoft Sentinel / Log Analytics: AWSCloudTrail (Amazon Web Services S3 connector)

let recon_apis = dynamic(["ListFoundationModels", "GetFoundationModelAvailability", "ListFoundationModelAgreementOffers",
                          "ListInferenceProfiles", "GetModelInvocationLoggingConfiguration", "GetUseCaseForModelAccess",
                          "InvokeModel", "InvokeModelWithResponseStream", "Converse", "ConverseStream"]);
let checks = dynamic(["GetModelInvocationLoggingConfiguration", "GetFoundationModelAvailability", "ListFoundationModelAgreementOffers"]);
AWSCloudTrail
| where TimeGenerated > ago(7d)
| where EventSource == "bedrock.amazonaws.com" and EventTypeName == "AwsApiCall"
| where EventName in (recon_apis)
| summarize DistinctApis = dcount(EventName), Apis = make_set(EventName), Calls = count(),
            Checks = countif(EventName in (checks))
    by UserIdentityArn, SourceIpAddress, HourBucket = bin(TimeGenerated, 1h)
| where DistinctApis >= 3 and Checks > 0
| order by DistinctApis desc, Calls desc

Splunk (SPL)

Splunk: AWS CloudTrail (Splunk Add-on for AWS, sourcetype aws:cloudtrail) · Set index= to the index that holds your CloudTrail data.

index=aws sourcetype=aws:cloudtrail eventSource="bedrock.amazonaws.com" eventType="AwsApiCall"
    eventName IN ("ListFoundationModels", "GetFoundationModelAvailability", "ListFoundationModelAgreementOffers",
                  "ListInferenceProfiles", "GetModelInvocationLoggingConfiguration", "GetUseCaseForModelAccess",
                  "InvokeModel", "InvokeModelWithResponseStream", "Converse", "ConverseStream")
    earliest=-7d
| eval is_check=if(in(eventName, "GetModelInvocationLoggingConfiguration", "GetFoundationModelAvailability",
                      "ListFoundationModelAgreementOffers"), 1, 0)
| bin _time span=1h
| stats dc(eventName) AS distinct_apis values(eventName) AS apis count AS calls sum(is_check) AS checks
        BY _time userIdentity.arn sourceIPAddress
| where distinct_apis >= 3 AND checks > 0
| sort - distinct_apis - calls

Expected false positives

  • Administrators exploring Bedrock in the console (console sessions also call these APIs)
  • Cloud security posture tools that inventory Bedrock settings

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.