ArtisanCloud PowerX through 1.0.2 contains a missing authorization vulnerability in the IAM role management API that allows authenticated tenant members to escalate privileges. Attackers holding even the lowest built-in role can call the permissions/grant-ids, permissions/grant, or bind/member routes to grant permissions or bind privileged roles, gaining tenant administrative access.
A time-of-check time-of-use (TOCTOU) race condition in the local port selection in AWS efs-utils (amazon-efs-utils) before 3.3.5 might allow a user that can start EFS mounts on a shared host to read and write a different Amazon EFS or S3 Files file system that is mounted on that same host, under that mount's IAM identity and access point, via concurrent mount operations. Only file systems mounted on the same host at the same time can be affected. Every mount involved was configured by the owner of that host, using credentials and access points that owner controls, against file systems whose owners had already granted that host access, so the issue grants no access beyond what the host was already permitted and does not cross AWS account boundaries unless the host's owner configured a cross-account mount. The practical risk is lateral access between workloads on the same host at the same time, for example one container reading another container's file system under that container's identity. The issue does not expose the other mount's IAM credentials or TLS session keys.
To remediate this issue, users should upgrade amazon-efs-utils to version 3.3.5 or later. Users of the Amazon EFS CSI Driver should upgrade to version 3.5.4 or later, and users of the Amazon EKS EFS CSI add-on should update to v3.5.4-eksbuild.1 or later.
ARTEX is an AI-native penetration framework observed in October 2026 targeting AWS environments, focusing on financial data. It uses autonomous LLM agents to plan and execute attacks, exploiting compromised long-term IAM access keys to search CloudWatch logs for sensitive information. The attackers also abuse Amazon Cognito user pools via anonymous sign-ups and token collection. Hundreds of exposed ARTEX consoles were found, mainly in Mandarin and hosted in Singapore and Hong Kong. The framework is linked to the China-aligned threat actor Storm-0940 and features advanced post-exploitation capabilities such as automated resource discovery and credential harvesting across cloud environments.
Between October 14 and October 26, 2026, a threat actor exploited IAM credentials from Qubole's AWS account to assume integration roles in customer AWS environments. The attacker used multiple IAM users and VPN addresses to conduct extensive reconnaissance, including enumerating storage, databases, and identity resources. Credentials were harvested from Lambda environment variables, Secrets Manager, and SSM Parameter Store, then validated with TruffleHog. Data was opportunistically exfiltrated from S3 buckets accessible to the assumed roles. A suspicious bootable Linux ISO image intended for remote deployment and system reconnaissance was also identified. The campaign exploited trusted cross-account relationships that remained active despite many integrations being unused and appears opportunistic rather than state-sponsored.
Dify versions before 1.15.0 have a server-side request forgery (SSRF) vulnerability that allows authenticated console users to bypass SSRF protections and send arbitrary HTTP requests to internal services and cloud metadata endpoints. This is due to insufficient Squid proxy configuration and lack of destination IP validation in the make_request() helper. Attackers can exploit various endpoints and workflow nodes to access internal components, including sandbox environments, internal APIs, PostgreSQL databases, and on AWS, the instance metadata service to read IAM credentials. The HTTP Request node also permits modifying internal services via POST, PUT, and DELETE methods. The vulnerability is fixed in version 1.15.0.
The npm package @ifasn/ms-doc-mcp, advertised as a Markdown toolkit, contains malicious code that probes the Alibaba Cloud instance metadata service to harvest temporary IAM credentials and other sensitive instance metadata. This harvested data is output to stdout and a local file, potentially exposing credentials beyond the local host. The behavior is unrelated to the package's declared functionality and poses a risk of credential leakage.
The @ifasn/mmd-doc-mcp npm package contains malicious code that probes cloud instance metadata service endpoints for AWS, GCP, Azure, and Alibaba clouds. It collects HTTP status codes from these endpoints to identify the cloud provider and presence of attached IAM/RAM roles, then sends this data in plaintext to a hardcoded external server. The package falsely advertises itself as a markdown table-of-contents server, but performs unauthorized reconnaissance of cloud credentials metadata services.
LLMJacking is a tactic where threat actors hijack enterprise access to Large Language Models (LLMs) by exploiting exposed cloud credentials and API keys. This activity has impacted organizations in North America, South America, and Asia Pacific across multiple sectors including financial services, retail, technology, and transportation. Unlike cryptojacking, LLMJacking causes sudden and massive spikes in fraudulent cloud usage charges, sometimes exceeding $100,000 USD daily and accumulating to over $1 million USD. Attackers gain initial access through stolen or exposed cloud identity tokens, IAM credentials, service account keys, and API keys, then escalate privileges and conduct reconnaissance. They monetize compromised resources via LLM Mixers, which are illicit routing gateways that proxy and resell unauthorized LLM access. The threat ecosystem is highly organized with specialized operational clusters coordinating both compromises and monetization infrastructure.
CVE-2026-119085 is a high-severity vulnerability in the publishing component of Amazon Virtual Engineering Workbench (VEW) caused by improper neutralization of special elements in its template engine. A remote authenticated attacker can exploit this flaw by submitting a crafted CloudFormation template containing malicious template engine directives via the product versioning API. Successful exploitation allows arbitrary code execution with the permissions of the publishing Lambda's IAM role, potentially exposing data and resources across all tenant projects. AWS has provided a fix in a newer version of the software and recommends redeploying from the latest version of awslabs/virtual-engineering-workbench to remediate the issue.
Insertion of sensitive information into a log file in Amazon Kinesis Producer Library for Java before 1.0.8, when the library's logger is configured at DEBUG log level, might allow local users to obtain the AWS credentials that the library resolved for the application. Separately, affected versions of the library pass these credentials to the native child process as command-line arguments, where a local user can read them from the process argument list.
To remediate this issue, users should upgrade to version 1.0.8 or later, rotate any long-term IAM user keys, revoke active role sessions where needed, and restrict or redact log data written by an affected version. 0.x users should migrate to software.amazon.kinesis:amazon-kinesis-producer 1.0.8 or later.
CVE-2026-61730 is a critical vulnerability in socfortress CoPilot versions prior to 0.1.77. It allows unauthenticated access to the /scoutsuite-report static files directory, exposing sensitive cloud security reports and potentially private Google service-account JSON keys during scans. This exposure can lead to unauthorized access to AWS, Azure, and GCP IAM principals and resource inventories. The issue is fixed in version 0.1.77.
Prometheus versions up to 3.15.1 have a credential leak vulnerability that enables attackers to capture Azure AD or Google IAM bearer tokens. This occurs when remote_write requests are redirected to a malicious host, allowing token replay and unauthorized metric writes to the victim's backend.
A flaw was found in KServe. A namespace-scoped user with create/update permissions on InferenceService custom resources can exploit this vulnerability by providing a malicious `storageUri`. This Server-Side Request Forgery (SSRF) vulnerability allows the `storage-initializer` init container to issue arbitrary outbound HTTP/HTTPS requests from inside the cluster. This can lead to information disclosure, including credential theft for the node's Identity and Access Management (IAM) role or access to internal cluster services.
CVE-2026-18061 is an XML External Entity (XXE) vulnerability in the AWS Advanced JDBC Wrapper's optional RemoteQueryCachePlugin. This plugin, when enabled with a shared cache, allows an attacker with write access to insert malicious XML into cached column values. When the application reads these cached results as XML, external entities are resolved, potentially exposing sensitive files accessible to the application process, including database and IAM role credentials. The vulnerability affects versions from 3.3.0 through 4.2.0 and is fixed in version 4.3.0. The plugin is not enabled by default. Mitigation includes disabling the plugin or restricting write access to the shared cache.
CVE-2026-89049 is a server-side request forgery (SSRF) vulnerability in the AWS Systems Manager Agent's Session Manager port forwarding feature. An authenticated user with port-forwarding permission can bypass the remote destination denylist due to improper validation of equivalent address representations. This allows the user to reach link-local endpoints and potentially obtain the managed instance's temporary IAM role credentials, enabling actions with those permissions from outside the instance. The vulnerability affects all versions prior to 3.3.4851.0. AWS has released version 3.3.4851.0 to fix this issue. Until upgrading, it is recommended to restrict the use of the AWS-StartPortForwardingSessionToRemoteHost document by scoping IAM permissions.