Injection Archives

AppWizard

May 21, 2025

Google says Gemini 2.5 models are only getting better with Deep Think and Flash

Google has announced enhancements to its Gemini 2.5 models, including the 2.5 Pro version and the new 2.5 Flash model, which improves speed and efficiency. The 2.5 Pro will include native audio output controls for developers to customize speech. Enhanced security measures will protect against malicious instructions and prompt injection attacks. Project Mariner's functionality will be integrated into Gemini and Vertex AI. Google is also introducing insightful summaries for developers to aid in debugging, along with cost control features through a "thinking budget." A generally available model will be released, and support for Model Context Protocol (MCP) will facilitate the integration of open-source tools into Gemini projects.

Winsage

May 20, 2025

Securing the Model Context Protocol: Building a safer agentic future on Windows

The Model Context Protocol (MCP) is a lightweight, open protocol functioning as JSON-RPC over HTTP, facilitating standardized discovery and invocation of tools. MCP defines three roles: MCP Hosts (applications accessing capabilities), MCP Clients (initiators of requests), and MCP Servers (services exposing functionalities). Windows 11 will incorporate MCP to enable developers to create intelligent applications leveraging generative AI. An early preview of MCP capabilities will be available for developer feedback. MCP introduces security risks, including cross-prompt injection, authentication gaps, credential leakage, tool poisoning, lack of containment, limited security review, registry risks, and command injection. To address these, Windows 11's MCP Security Architecture will establish security requirements for MCP servers, ensuring user safety and transparency, enforcing least privilege, and implementing security controls like proxy-mediated communication, tool-level authorization, a central server registry, and runtime isolation. MCP servers must comply with security requirements, including mandatory code signing, unchanged tool definitions at runtime, security testing, mandatory package identity, and declared privileges. An early private preview of MCP server capability will be offered to developers post-Microsoft Build for feedback, with a secure-by-default enforcement strategy planned for broader availability. Microsoft aims to enhance defenses continuously and collaborate with partners to bolster MCP's security framework.

Winsage

May 12, 2025

Microsoft Defender Discovers Zero-Day Vulnerability in Windows CLFS

The deployment of PipeMagic preceded a sophisticated exploit targeting the Common Log File System (CLFS) kernel driver, initiated from a dllhost.exe process. The exploit began with the NtQuerySystemInformation API, which leaked kernel addresses to user mode. In Windows 11, version 24H2, access to specific System Information Classes within this API was restricted to users with SeDebugPrivilege, rendering the exploit ineffective on this version. The exploit then used a memory corruption technique with the RtlSetAllBits API to overwrite its process token with 0xFFFFFFFF, granting it all available privileges and enabling process injection into SYSTEM-level operations. A CLFS BLF file was created at C:ProgramDataSkyPDFPDUDrv.blf, marking the exploit's activity.

Winsage

April 22, 2025

Critical Windows Update Stack Vulnerability Allows Code Execution & Privilege Escalation

A security vulnerability identified as CVE-2025-21204 has been discovered in the Windows Update Stack, allowing local attackers to execute unauthorized code and escalate privileges to SYSTEM-level access. This vulnerability, with a CVSS score of 7.8 (High), affects Windows 10 versions 1507, 1607, and 1809, among likely other supported Windows 10/11 and Windows Server versions. The flaw arises from a design issue where Windows Update processes do not properly follow directory junctions, enabling attackers with limited user privileges to redirect trusted paths to locations containing malicious code. Microsoft has introduced a mitigation strategy in its April 2025 cumulative update, which includes creating a new folder at the root of system drives and implementing detection rules for suspicious junction creations. Organizations are advised to apply the April 2025 security updates, restrict ACLs on specific directories, prevent symbolic link creation, and monitor file creation activities in certain directories.

Tech Optimizer

April 15, 2025

New ‘Waiting Thread Hijacking’ Malware Technique Evades Modern Security Measures

Security researchers have developed a new malware process injection technique called "Waiting Thread Hijacking" (WTH), which executes harmful code within legitimate processes while avoiding detection by security measures. This method improves upon traditional Thread Execution Hijacking by using a different sequence of operations that bypasses commonly monitored API calls. WTH involves allocating memory and injecting malicious payloads using standard functions, identifying dormant threads within the target process, acquiring thread context with less suspicious permissions, and overwriting the return address on the stack with the injected shellcode. The technique ensures stability by preserving the original state of the thread and allows it to resume normal operations after executing the malicious code. Additionally, WTH employs an obfuscation technique that distributes its steps across multiple child processes to evade behavioral detection systems. While WTH can avoid many conventional detection triggers, it is not completely immune, as some Endpoint Detection and Response (EDR) solutions can block unauthorized memory writes. Check Point Research has observed that WTH is effective against certain EDRs while others can block it but not older methods, illustrating the variability in EDR capabilities.

Winsage

April 9, 2025

Exploitation of CLFS zero-day leads to ransomware activity

The Microsoft Threat Intelligence Center (MSTIC) and Microsoft Security Response Center (MSRC) identified a zero-day elevation of privilege vulnerability in the Windows Common Log File System (CLFS), designated as CVE-2025-29824. This vulnerability has been exploited by the PipeMagic malware, attributed to the group Storm-2460, affecting organizations in various sectors across the U.S., Venezuela, Spain, and Saudi Arabia. Microsoft released security updates on April 8, 2025, to address this vulnerability. The exploit allows attackers with standard user privileges to escalate their access, and it was executed from a dllhost.exe process. The exploit utilizes memory corruption techniques and the RtlSetAllBits API to gain all privileges and inject processes into SYSTEM processes. Post-exploitation, the attackers injected a payload into winlogon.exe, leading to ransomware activity characterized by file encryption and the creation of ransom notes. Indicators of compromise include the creation of a CLFS BLF file at C:ProgramDataSkyPDFPDUDrv.blf, command lines associated with ransomware activities, and specific domains linked to PipeMagic. Microsoft advises applying security updates promptly and recommends strategies for mitigating ransomware threats, including enabling cloud-delivered protection and utilizing device discovery.

Tech Optimizer

April 3, 2025

1,500+ PostgreSQL Servers Compromised With Fileless Malware Attack

A cryptojacking campaign has targeted over 1,500 organizations by exploiting inadequately secured PostgreSQL database servers. The attackers, identified as JINX-0126, use advanced techniques including credential brute-forcing and fileless execution to deploy Monero (XMR)-mining malware. Approximately 30% of cloud-hosted PostgreSQL servers are affected due to weak or default credentials. The attackers execute commands using PostgreSQL’s COPY FROM PROGRAM function, bypassing standard detection mechanisms. They have been linked to three cryptocurrency wallets with around 550 active mining workers, generating a hashrate of 4.04 GH/s and approximately €10.40 per hour in XMR revenue. The attack begins with credential spraying against default accounts, followed by an SQL injection to fetch the payload. The malware operates in memory, modifies PostgreSQL’s configuration to maintain persistence, and creates cron jobs for reactivation. The campaign reveals significant security gaps in cloud environments, with recommendations for improved access controls and monitoring.

Tech Optimizer

March 5, 2025

Digest #162: IBM Acquires HashiCorp, No TCP/UDP, GitHub Data Theft, FinOps Future, Postgres Graph, AWS Config, Airflow, Kubernetes, Nginx, Terraform & More!

IBM has finalized its acquisition of HashiCorp. Concerns have emerged about the security of GitHub modifications, with reports indicating that thousands may compromise user data. A recent court case involving SQL injection highlights the importance of robust security measures in database management. CPU usage has surged by 3,200%. The tutorials section includes guides on utilizing Postgres as a graph database, automating cost checks with AWS Config, reducing AWS Glue expenses, navigating Kubernetes audit log challenges, mastering slicing SLOs, implementing advanced Nginx hardening techniques, exploring Amazon EKS auto mode with Terraform, leveraging GitHub Copilot for Azure DevOps, and enhancing infrastructure provisioning skills. Open-source projects highlighted include Yaak, Telescope, AtomixDB, PG Capture, TFBuddy, Globstar, and Tach.

Winsage

March 4, 2025

CISA Alerts on Active Exploitation of Cisco Small Business Router Flaw

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has issued a critical alert about a command injection vulnerability (CVE-2023-20118) affecting Cisco Small Business RV Series Routers, which are end-of-life. This vulnerability, rated 6.5 on the CVSSv3.1 scale, allows authenticated attackers to execute arbitrary commands with root privileges. The affected models include RV016, RV042, RV042G, RV082, RV320, and RV325, running firmware versions released before April 2023. Cisco will not provide patches for these devices. CISA mandates that federal agencies either implement mitigations or stop using the routers by March 24, 2025. Private organizations are also encouraged to address the issue, especially due to exploitation attempts linked to the PolarEdge botnet campaign. Administrators are advised to restrict administrative access, monitor logs for unusual activity, and consider decommissioning affected devices. The continued use of unpatched routers poses significant risks to critical infrastructure, particularly in small business and remote work environments.

Winsage

March 4, 2025

U.S. CISA adds Multiple Cisco Small Business RV Series Routers, Hitachi Vantara Pentaho BA Server, Microsoft Windows Win32k, and Progress WhatsUp Gold flaws to its Known Exploited Vulnerabilities catalog

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has expanded its Known Exploited Vulnerabilities (KEV) catalog, which now includes several significant security flaws: - CVE-2023-20118: A command injection vulnerability in Cisco Small Business RV Series Routers with a CVSS score of 6.5, allowing authenticated remote attackers to execute arbitrary commands. Cisco will not provide a fix for this issue. - CVE-2022-43939: An authorization bypass vulnerability in the Hitachi Vantara Pentaho BA Server. - CVE-2022-43769: A special element injection vulnerability in the Hitachi Vantara Pentaho BA Server. - CVE-2018-8639: An elevation of privilege vulnerability in Microsoft Windows with a CVSS score of 7.8, allowing an attacker to run arbitrary code in kernel mode. - CVE-2024-4885: An unauthenticated remote code execution vulnerability in Progress WhatsUp Gold with a CVSS score of 9.8, allowing command execution with iisapppoolnmconsole privileges. CISA has mandated that federal agencies address these vulnerabilities by March 24, 2025, under Binding Operational Directive (BOD) 22-01, and advises private organizations to review the KEV catalog for necessary actions.