Exposed AWS keys mass-tested for Bedrock LLMjacking
Datadog observed automated, internet-scale testing of exposed AWS keys specifically for Amazon Bedrock model access, a reconnaissance step toward LLMjacking.
What happened
Datadog documents automated testing of exposed AWS keys specifically for Amazon Bedrock access (ListFoundationModels / Invoke / Converse patterns), showing LLMjacking reconnaissance at internet scale beyond generic AWS abuse.
How it happened
Read the flow as text & sources
A catalog reconstruction of the supplied source account. Unknown steps remain explicit; an edge describes the reported relationship between steps.
- 01 · Exposed AWS keys are collected
Attackers obtain credentials exposed outside the legitimate account boundary.
Source for step 1 - 02 · Automated key tests run
Requests check model listing, invocation and conversation permissions.
Hunt: A key enumerates Bedrock models from an unfamiliar origin, followed by unexpected model access or usage.
The starter queries cover discovery only. Do not assume default CloudTrail management logs contain InvokeModel / Converse requests or prompt content.
Source for step 2 - 03 · Bedrock access is identified
Testing distinguishes keys that can use model inference services.
Hunt: A key enumerates Bedrock models from an unfamiliar origin, followed by unexpected model access or usage.
The starter queries cover discovery only. Do not assume default CloudTrail management logs contain InvokeModel / Converse requests or prompt content.
Source for step 3 - 04 · Per-key abuse is not quantified
The supplied entry establishes reconnaissance at scale, rather than a bill or victim count for every key.
Source for step 4
Why it matters
Reported impact: the cited report does not establish a confirmed victim outcome.
Check these source-linked conditions against your setup.
The campaign tests exposed AWS keys for Bedrock access.
Conditions to check- Confirm that the credentials belong to AWS and have Bedrock permissions; a product subscription alone is not the condition.
Source-linked applicability guidance. Confirm current scope in the original report.
What you can do
Protection & prevention
Inventory Bedrock-enabled keys, rotate exposed credentials and correlate discovery calls with actual model usage.
Secret scanning; CloudTrail detections for Bedrock enum+invoke from unexpected keys
Catalog defensive guidance based on the reported behavior. Confirm scope and current mitigations in the original source.
Detection queries
Draft query starters grouped by the attack stages where they apply.
Start with the logs you collect. Each hunt identifies its required visibility, field mapping and limits. Queries are catalog-authored drafts using reusable behavior profiles; test with known activity and your normal baseline before alerting.
02Automated key tests runBedrock discovery / key testingNative logs · Draft
Requests check model listing, invocation and conversation permissions.
What to look for
A key enumerates Bedrock models from an unfamiliar origin, followed by unexpected model access or usage.
Visibility: Defender-side logs where collected
Required logs
- AWS CloudTrail management events
- Bedrock invocation logging / relevant data events when explicitly enabled
- Key inventory and inference usage
Coverage limit. The starter queries cover discovery only. Do not assume default CloudTrail management logs contain InvokeModel / Converse requests or prompt content.
AWS discovery candidates: Sigma CloudTrail, Sentinel AWSCloudTrail, and Splunk aws:cloudtrail. Invocation logging is a separate prerequisite.
Sigma
title: Hunt - Bedrock discovery / key testing
id: 89e0e836-0c7b-44b4-a9d4-3524f68318b2
status: experimental
description: A key enumerates Bedrock models from an unfamiliar origin, followed
by unexpected model access or usage.
author: MagicSword OpenAttacks catalog
date: 2026-10-23
falsepositives:
- Authorized activity matching this broad hunt; baseline and investigate
before alerting.
level: low
logsource:
product: aws
service: cloudtrail
detection:
selection:
eventSource: bedrock.amazonaws.com
eventName:
- ListFoundationModels
- GetFoundationModel
condition: selection
KQL
// Discovery only; correlate with principal inventory and enabled invocation logs.
AWSCloudTrail
| where TimeGenerated > ago(7d)
| where EventSource == "bedrock.amazonaws.com"
| where EventName in ("ListFoundationModels", "GetFoundationModel")
| summarize Requests=count() by UserIdentityArn, SourceIpAddress, UserAgent, EventName, bin(TimeGenerated, 1h)
| order by Requests descSPL
sourcetype="aws:cloudtrail" eventSource="bedrock.amazonaws.com"
eventName IN ("ListFoundationModels", "GetFoundationModel")
| bin _time span=1h
| stats count by _time userIdentity.arn sourceIPAddress userAgent eventName
| sort - countInvestigation steps
- Group ListFoundationModels / GetFoundationModel calls by key, source IP and user agent; compare against normal application identities.
- Correlate with invocation logs and cost. Discovery alone does not establish a stolen key or successful inference.
03Bedrock access is identifiedBedrock discovery / key testingNative logs · Draft
Testing distinguishes keys that can use model inference services.
What to look for
A key enumerates Bedrock models from an unfamiliar origin, followed by unexpected model access or usage.
Visibility: Defender-side logs where collected
Required logs
- AWS CloudTrail management events
- Bedrock invocation logging / relevant data events when explicitly enabled
- Key inventory and inference usage
Coverage limit. The starter queries cover discovery only. Do not assume default CloudTrail management logs contain InvokeModel / Converse requests or prompt content.
AWS discovery candidates: Sigma CloudTrail, Sentinel AWSCloudTrail, and Splunk aws:cloudtrail. Invocation logging is a separate prerequisite.
Sigma
title: Hunt - Bedrock discovery / key testing
id: 89e0e836-0c7b-44b4-a9d4-3524f68318b2
status: experimental
description: A key enumerates Bedrock models from an unfamiliar origin, followed
by unexpected model access or usage.
author: MagicSword OpenAttacks catalog
date: 2026-10-23
falsepositives:
- Authorized activity matching this broad hunt; baseline and investigate
before alerting.
level: low
logsource:
product: aws
service: cloudtrail
detection:
selection:
eventSource: bedrock.amazonaws.com
eventName:
- ListFoundationModels
- GetFoundationModel
condition: selection
KQL
// Discovery only; correlate with principal inventory and enabled invocation logs.
AWSCloudTrail
| where TimeGenerated > ago(7d)
| where EventSource == "bedrock.amazonaws.com"
| where EventName in ("ListFoundationModels", "GetFoundationModel")
| summarize Requests=count() by UserIdentityArn, SourceIpAddress, UserAgent, EventName, bin(TimeGenerated, 1h)
| order by Requests descSPL
sourcetype="aws:cloudtrail" eventSource="bedrock.amazonaws.com"
eventName IN ("ListFoundationModels", "GetFoundationModel")
| bin _time span=1h
| stats count by _time userIdentity.arn sourceIPAddress userAgent eventName
| sort - countInvestigation steps
- Group ListFoundationModels / GetFoundationModel calls by key, source IP and user agent; compare against normal application identities.
- Correlate with invocation logs and cost. Discovery alone does not establish a stolen key or successful inference.
Technical detailsTooling, impact & MITRE ATLAS
Attack context
- AI's role
- AI credential theft
Tooling as reported
AI products, models and artifacts
Other tooling
- Stolen AKIA keys
Detection rules
- Review attack-specific indicatorsGUIDANCE
Monitor for the techniques described in Exposed AWS keys mass-tested for Bedrock LLMjacking. Refer to sources for specific IOCs and TTPs.
Source defensive notes
Secret scanning; CloudTrail detections for Bedrock enum+invoke from unexpected keys
MITRE ATLAS
AI-specific catalog annotations based on the reported behavior; these are not MITRE case-study assignments.
Exposed AWS credentials are tested for model-service access. Source
Sources & research credit
1 sourceResearch by Datadog Security Labs · Datadog
Record & downloads
- Source record name
- Exposed AWS keys mass-tested for Bedrock LLMjacking
- Entry ID
- datadog-bedrock-key-testing-2025
- Catalog review
- October 20, 2026
- Category
- attacks-on-ai
- AI involvement confidence
- confirmed