Follow the evidence.

AWS AppFlow secrets disclosure

AppFlow had an undocumented service called sandstoneconfigurationservicelambda. An undocumented field (awsOwnedManagedAppCredentialsArn) could be used during connector registration and connector updates. Specifying a victim's Secret ARN as that field disclosed the clientId and clientSecret, so long as the victim Secret ARN belonged to a connection profile which is of the type OAuth or contains cli

Confused Deputy Vulnerability in Amazon DataZone

A vulnerability in Amazon DataZone allowed potential attackers to assume roles in AWS accounts by exploiting a confused deputy problem. This could have granted unauthorized access to sensitive data managed by DataZone or other AWS services accessible by the IAM role trusting DataZone. The issue has been resolved, with no customers reportedly impacted.

AWS EKS Logged ServiceAccount Tokens in Plaintext

AWS EKS was logging ServiceAccount tokens in plaintext, including those used for AssumeRoleWithWebIdentity and connecting to the Kubernetes API server. This issue affected clusters between March 2020 and May 2021, potentially exposing sensitive credentials in CloudWatch logs.

Google Cloud GKE Unsecure Sys:All Binding

The system:authenticated group in Kubernetes is a special group that includes all authenticated entities, including human users and service accounts. Anyone who successfully authenticates to the Kubernetes API server, regardless of the authentication method used, will be automatically included in this unique group. Thus, it will share the same roles and permissions of the group. This misunderstand

Abusing Service Tags to Bypass Azure Firewall Rules

Tenable Research discovered a vulnerability in Azure allowing attackers to bypass firewall rules based on Service Tags by forging requests from trusted services. It affects over 10 Azure services and enables access to internal/private Azure resources. Microsoft updated documentation to clarify Service Tags' security limitations.

ModeLeak: LLM Model Exfiltration Vulnerability in Vertex AI

A vulnerability in GCP's Vertex AI service allows privilege escalation and unauthorized access to sensitive LLM models. Attackers can exfiltrate these models by exploiting misconfigurations in access controls and service bindings. By exploiting custom job permissions, researchers were able to escalate their privileges and gain unauthorized access to all data services in the project. In addition, d

Azure Machine Learning SSRF

Certain API endpoints on ml.azure.com and ai.azure.com used for adding/viewing data connections could be leveraged for server side request forgeries (SSRF). While they do have protections to restrict making requests to internal hosts, it was possible to circumvent those protections using a 301 or 302 redirect response which points to a sensitive host.

Issue with AWS Deployment Framework

CVE-2024-37293 affects the AWS Deployment Framework's bootstrap process, potentially allowing privilege escalation if an actor has permissions to change CodeBuild projects or Lambda functions. The issue is fixed in version 4.0 and above. AWS recommends immediate upgrade and temporary mitigation by adding a permissions boundary to roles created by ADF in the management account.

Poisoning GitHub's Runner Images Supply Chain Attack

A critical vulnerability in GitHub's actions/runner-images repository allowed arbitrary code execution on self-hosted runners, potentially enabling modification of GitHub's runner base images. The flaw stemmed from misconfigured self-hosted runners on a public repository with default workflow approval settings. The researcher gained persistence, accessed secrets, and could have inserted malicious

Bucket Monopoly Attack on AWS Services

Researchers discovered critical vulnerabilities in 6 AWS services that could allow attackers to breach accounts through malicious S3 buckets. By claiming predictable bucket names, attackers could inject code, steal data, or gain admin access. AWS has since fixed the issues, but the attack vector may still apply to other services and open source projects.

GitHub Copilot Chat Vulnerable to Data Exfiltration

GitHub Copilot Chat VS Code Extension was vulnerable to data exfiltration via prompt injection when analyzing untrusted source code. The vulnerability allowed attackers to access previous conversation turns and append information from the chat history to an image URL, which was then automatically retrieved by Copilot, sending the data to the attacker.

Entra ID Bug Creates Immutable Users

A bug in Entra ID restricted management administrative units allowed creating immutable users that couldn't be modified or disabled, even by Global Administrators. This could enable an attacker to protect a compromised account from containment. The issue was caused by a timing vulnerability when removing users from restricted AUs and required specific steps to remediate affected accounts.

GraphNinja

A vulnerability in Microsoft Graph allowed attackers to conduct password-spray attacks without detection. The issue involved switching the 'common' authentication endpoint with that of an unrelated tenant, thereby avoiding the appearance of logon attempts in the victim's logs. This technique could allow attackers to validate user credentials through verbose error messages, but actual successful lo

Amazon Q for Business Data Exfiltration

An Indirect Prompt Injection attack can cause the LLM to return markdown tags. This allows an adversary who’s data makes it into the chat context (e.g via an uploaded file) to achieve data exfiltration of the victim’s data by rendering hyperlinks.

Subdomain Takeover Vulnerability in GitLab Pages

A vulnerability in GitLab Pages allowed attackers to take over dangling custom domains pointing to 'instanceX.gitlab.io'. The issue occured when adding an unverified custom domain to GitLab Pages, which serves content for 7 days before disabling. This could lead to cookie stealing, phishing campaigns, and bypassing of Content-Security Policies and CORS.

CloudImposer

Google Cloud Composer is a managed service for Apache Airflow. Tenable discovered that the Cloud Composer package was vulnerable to dependency confusion, which could have allowed attackers to inject malicious code when the package was compiled from source. This could have led to remote code execution on machines running Cloud Composer, which include various other GCP services as well as internal s

AWS Client VPN buffer overflow

The AWS Client VPN service was found to be affected by two vulnerabilities which could potentially allow malicious actors with access to a user’s device to execute arbitrary commands with elevated privileges, including escalating to root access. Both vulnerabilities stem from buffer overflow issues, a common programming error that can be exploited to overwrite memory and gain unauthorized control

Critical GitLab Account Takeover Vulnerability

GitLab addressed a critical vulnerability, CVE-2023-7028, affecting managed SaaS gitlab.com instance as well as self-hosted versions 16.1 to 16.7.1. The flaw could allow account takeovers via unverified email password resets. Third party could intercept the password reset request, add their own email to the request and forward it. GitLab would then send the reset link to the added 3rd-party email.

Google Workspace Domain-Wide Delegation Flaw

Unit 42 researchers discovered a security risk in Google Workspace's domain-wide delegation feature that allows a GCP identity with necessary permissions to generate access tokens to impersonate Google Workspace users and access their data. This mismatch between GCP permissions and Google Workspace access could be exploited by malicious insiders or attackers with stolen credentials.

Abusing AWS Serverless Image Handler Configuration Weakness

AWS solution 'Dynamic Image Transformation for Amazon CloudFront', prior to version 6.2.6, contains a configuration weakness. The Lambda role doesn't constrain bucket access, and the environment variable can be set to a wildcard, allowing access to any bucket. This could potentially lead to unintended access to sensitive images across multiple buckets in the AWS account.

Code Execution in Azure API Management Developer Portal

A vulnerability in Azure API Management Developer Portal allows arbitrary code execution and secret exfiltration. The issue stems from a workflow that loads untrusted data from opened issues, potentially allowing attackers to inject malicious commands. This could lead to code execution in the runner, granting access to sensitive tokens and permissions.

Copilot Studio information disclosure via SSRF

Google Cloud ConfusedComposer Privilege Escalation Vulnerability

Tenable discovered a privilege escalation vulnerability in Google Cloud Platform's Cloud Composer service, dubbed ConfusedComposer. It allowed users with composer.environments.update permission to escalate privileges to the default Cloud Build service account by injecting malicious PyPI packages. This could grant broad permissions across the victim's GCP project.

CloudWatch Dashboard Sharing Exposes EC2 Tags

A vulnerability in AWS CloudWatch dashboard sharing allowed viewers to access EC2 instance tags and potentially invoke Lambda functions in the source account. The issue stemmed from a logic bug in the AWS Console combined with a "fail open" condition in Amazon Cognito. AWS has since patched the vulnerability.