QCS Security Advisory Desk
Security Advisory Desk.
Monitor network-edge vulnerabilities, known exploitation, affected products, mitigations, and vendor patch guidance without waiting for a weekly editorial cycle.
- 01Official sourceMonitor vendors and trusted authorities.
- 02Verify contextCheck products, exposure, and evidence.
- 03Set prioritySeparate urgent action from useful reading.
- 04Act or learnPatch, mitigate, investigate, or prepare.
Live intelligence
See the items that can change today's patch or mitigation plan.
Priority combines source severity, known exploitation, remote attack conditions, recency, and network-edge relevance.
Amazon Web Services / Priority 70
CVE-2026-104020 - Uncontrolled recursion in the Ion reader in Amazon Ion Python
Amazon Ion Python is a software library that helps applications read and write data in the Amazon Ion format. A security issue was found where maliciously crafted data with deep nested structures can cause the Ion reader to crash or raise errors. This leads to the application being unable to function properly, a condition known as denial of service. To address this, users should upgrade to Amazon Ion Python version 0.15.0 or later, which includes a fix. Alternatively, users can disable a specific C extension to help mitigate the problem.
- Published
- 1 Oct 2026
- CVSS
- Not scored
- Verified
- 5 Oct 2026
Amazon Web Services / Priority 70
CVE-2026-97662 - Argument injection in AWS security-agent-mcp-server diff scan
AWS identified a security issue in security-agent-mcp-server, an open-source tool used by AI assistants for scanning source code. A specially crafted input in the differential scan function can be mistakenly treated as a command option, allowing an attacker to write or overwrite files outside the intended area, circumventing protections. AWS fixed this in version 0.2.0 and advises upgrading to this version to secure your environment.
- Published
- 1 Oct 2026
- CVSS
- Not scored
- Verified
- 5 Oct 2026
Amazon Web Services / Priority 70
CVE-2026-100308 - GluonTS arbitrary command execution during model deserialization
A security vulnerability (CVE-2026-100308) was found in Amazon GluonTS, a library for deep learning models on time series data. It allows attackers to run any operating system commands if they load a manipulated model from an untrusted source. This can happen when users call deserialization functions on model directories they don't fully control.
- Published
- 29 Sep 2026
- CVSS
- Not scored
- Verified
- 5 Oct 2026
Google Cloud / Priority 70
GCP-2026-059
In the HTTP Connector component of Integration Connectors used in Google Cloud, there was a missing authorization check vulnerability discovered. This flaw allowed an authenticated attacker to attach an unauthorized service account, thereby escalating their privileges and potentially taking over an entire Google Cloud project. Google fixed this issue on December 11, 2025, and no customer action is required because the patch has already been applied.
- Published
- 28 Sep 2026
- CVSS
- Not scored
- Verified
- 3 Oct 2026
Google Cloud / Priority 70
GCP-2026-060
A security issue was found in Google's Slurm sbcast tool that could allow shared library files to bypass security checks and potentially cause cluster nodes to crash. This affects the stability and security of computing clusters using this tool. Users should review the linked Cluster Toolkit and Cluster Director security bulletins for guidance.
- Published
- 28 Sep 2026
- CVSS
- Not scored
- Verified
- 3 Oct 2026
Google Cloud / Priority 70
GCP-2026-061
A security flaw in containerd's CRI implementation allows containers restored from untrusted checkpoints to bypass destination security settings and gain elevated privileges. Default Google Kubernetes Engine (GKE) clusters are not affected because they do not include the necessary tool. No immediate action is required for default clusters, but if custom software including the checkpoint restore tool is installed, follow specific security guidance.
- Published
- 28 Sep 2026
- CVSS
- Not scored
- Verified
- 3 Oct 2026
Google Cloud / Priority 70
GCP-2026-062
Google Cloud discovered multiple security vulnerabilities in Slurm, a workload manager used in cluster computing. These flaws affect certain Cluster Toolkit blueprints that rely on specific image versions. The vulnerabilities involve the slurmstepd daemon, remote procedure call (RPC) request handling, and the accounting database. Users should consult the Cluster Toolkit security bulletin for instructions and details to address these issues.
- Published
- 28 Sep 2026
- CVSS
- Not scored
- Verified
- 3 Oct 2026
Google Cloud / Priority 70
GCP-2026-065
A security flaw was discovered in Google Cloud's Application Integration JavaScript Task, where an authenticated user with standard permissions could run harmful code on shared production servers. Google fixed this issue on June 28, 2026, and no action is required from customers.
- Published
- 28 Sep 2026
- CVSS
- Not scored
- Verified
- 3 Oct 2026
Google Cloud / Priority 70
GCP-2026-064
A security flaw was found in Google Cloud's Application Integration where an authenticated user could misuse internal-only tasks to perform unauthorized actions within Google's internal production network with elevated privileges. This issue was fixed by Google on June 17, 2026, and customers do not need to take any action.
- Published
- 28 Sep 2026
- CVSS
- Not scored
- Verified
- 3 Oct 2026
Google Cloud / Priority 70
GCP-2026-063
A security issue was found in Google Security Operations SOAR where a logged-in attacker could gain full system administrator rights by sending a specially crafted internal authentication header. Google fixed this issue in version 6.3.85 and automatically upgraded all customers to this fixed version. Customers do not need to take any action.
- Published
- 28 Sep 2026
- CVSS
- Not scored
- Verified
- 3 Oct 2026
Microsoft / Priority 70
CVE-2026-65660: Microsoft SharePoint Code Injection Vulnerability
Microsoft SharePoint has a security flaw that allows an attacker with authorized access to inject malicious code and potentially execute it over the network.
- Published
- 25 Sep 2026
- CVSS
- Not scored
- Verified
- 6 Oct 2026
Amazon Web Services / Priority 70
CVE-2026-94384 - Missing Authorization in AmazonConnectSalesforceLambda sfExecuteAWSService
Amazon's AmazonConnectSalesforceLambda application has a security issue in one of its setup functions called sfExecuteAWSService. This function lets authorized users pass commands to powerful AWS services without properly checking if they're allowed to do so. Because of this, someone with permission to run this function could perform actions they usually wouldn't have permission for, risking security. To fix this, users must upgrade the application, remove or disable the risky function after setup, or tightly control who can run it.
- Published
- 22 Sep 2026
- CVSS
- Not scored
- Verified
- 22 Sep 2026
