Tia N.List

Reported by 2 outlets

LLMjacking Attack Abuses Leaked AWS Credentials to Hijack Amazon Bedrock AI Models

hot0.0
Why this score
Breadth
1.00 / 2.0 · 2 sources
Authority
1.20 / 2.0 · best outlet weight
Severity
0.70 / 5.0 · CVSS + content signals
Velocity
0.00 / 2.0 · updates in 48h
Pickup speed
0.60 / 0.6 · time to second source
Recency
× 0.00 temporal multiplier
Community
+ 0.00 / 1.5 · reddit signal
Score
0.0 / 10 = base × recency + community
gbhackers.comcyberpress.org first seen 39d ago

Timeline

  1. original reporting

    LLMjacking Attack Abuses Leaked AWS Credentials to Hijack Amazon Bedrock AI Models

    gbhackers.com · 39d ago · September 4th, 2026

    Threat actors are increasingly converting stolen cloud credentials into access to costly generative AI services, a technique known as LLMjacking. FortiGuard Labs reported an incident involving Amazon Bedrock in which a leaked, long-lived AWS IAM access key with AdministratorAccess permissions was used to create a new identity, subscribe to foundation models, and run inference at […] The post LLMjacking Attack Abuses Leaked AWS Credentials to Hijack Amazon Bedrock AI Models appeared first on GBHackers Security | #1 Globally Trusted Cyber Security News Platform .

  2. update from cyberpress.org

    Hackers Abuse Leaked AWS IAM Keys to Hijack Amazon Bedrock AI Model Access

    cyberpress.org · 39d ago · September 4th, 2026

    A newly documented cloud intrusion demonstrates how attackers can transform a single exposed AWS credential into a lucrative AI abuse operation, hijacking access to premium Amazon Bedrock foundation models and billing the victim for inference costs. Security researchers at FortiGuard Labs described the technique as LLMjacking: the abuse of compromised cloud credentials to access and resell costly large-language-model services. Unlike cryptomining or data theft, the attacker monetizes the target organization’s existing relationship with its cloud provider. Hackers Abuse Leaked AWS IAM Keys The incident began with a long-lived AWS Identity and Access Management (IAM) access key that had AdministratorAccess permissions the broadest level of privilege available to an AWS identity. With the stolen key, the attacker could operate as a highly privileged account owner. Researchers observed the threat actor creating a new IAM user within the compromised AWS environment, providing a fresh identity that could be used for subsequent activity without continuously relying on the originally exposed key. Attack Chain (Source: FORTINET) The newly created identity was then used to subscribe to foundation models available through AWS Marketplace. The attackers issued CreateAgreementRequest and AcceptAgreementRequest calls to the AWS Marketplace agreement service to accept the commercial terms required to access premium models. Once the agreements were active, the attacker began invoking the mod

Covered in

digest 2026-09-04

Source on GitHub