Shai-Hulud’s ‘Second Coming’ npm Malware Infects Popular Developer Packages
Shai-Hulud malware campaign, npm Packages
Continue ReadingShai-Hulud malware campaign, npm Packages
Continue ReadingSummary : Fluent Bit is a widely used opensource tool for collecting and forwarding logs in cloud and containers like Kubernetes environments. A chain of 5 critical vulnerabilities discovered by Oligo Security team and findings reveal that attackers can misuse via Remote code execution putting cloud and container at risk.
| Severity | Critical |
| CVSS Score | 9.1 |
| CVEs | CVE-2025-12969, CVE-2025-12970, CVE-2025-12972, CVE-2025-12977, CVE-2025-12978 |
| POC Available | No |
| Actively Exploited | No |
| Exploited in Wild | No |
| Advisory Version | 1.0 |
Overview
These vulnerabilities are CVE-2025-12977 CVE-2025-12970, CVE-2025-12969, CVE-2025-12978 , CVE-2025-12972. The vulnerabilities allow attackers to bypass authentication, manipulate log routing, achieve remote code execution, potentially leading to full compromise of cloud and Kubernetes environments using Fluent Bit for logging and observability.
Organizations relying on Fluent Bit must upgrade to the fixed versions and harden configurations to prevent remote takeover and log tampering.
| Vulnerability Name | CVE ID | Product Affected | Severity | CVSS Score | Fixed Version |
| Fluent Bit Tag_Key Input Validation Bypass | CVE-2025-12977 | Fluent Bit | Critical | 9.1 | v4.0.12+ , v4.1.1+ , v4.2.0+ |
| Fluent Bit Docker Input Stack Buffer Overflow | CVE-2025-12970 | Fluent Bit | High | 8.8 | v4.0.12+ , v4.1.1+ , v4.2.0+ |
| Fluent Bit Forward Input Authentication Bypass | CVE-2025-12969 | Fluent Bit | Medium | 6.5 | v4.0.12+ , v4.1.1+ , v4.2.0+ |
| Fluent Bit Tag Spoofing via Partial Tag_Key Match | CVE-2025-12978 | Fluent Bit | Medium | 5.4 | v4.0.12+ , v4.1.1+ , v4.2.0+ |
| Fluent Bit File Output Path Traversal | CVE-2025-12972 | Fluent Bit | Medium | 5.3 | v4.0.12+ , v4.1.1+ , v4.2.0+ |
Technical Summary
Fluent Bit vulnerabilities center around unsafe handling of tags and inputs, enabling attackers to manipulate routing, file paths and memory in ways that directly impact host systems and downstream security tooling.
These flaws can allow path traversal and arbitrary file writes, which in many real-world setups may escalate to remote code execution and persistent node compromise.
Additional vulnerabilities include stack buffer overflows and missing authentication checks that let attackers crash agents, execute code and inject false telemetry into trusted logging pipelines.

Source: Oligo.security
| CVE ID | Vulnerability Details | Impact |
| CVE-2025-12977 | Improper input validation allows injection of control chars, newlines, and path traversal sequences in tag values. | Log corruption and output injection. |
| CVE-2025-12970 | Stack buffer overflow on container name copy due to lack of length check. | Crash or RCE. |
| CVE-2025-12969 | Authentication bypass disables user-based auth, allowing unauthenticated log injection. | Unauthorized log injection. |
| CVE-2025-12978 | Partial string comparison on Tag_Key lets attacker spoof tags by guessing first char. | Manipulation of log routing and filtering. |
| CVE-2025-12972 | Path traversal via unsanitized tags causes arbitrary file write and possible remote code execution. | Arbitrary file write and RCE. |
Remediation:
Here are some recommendations below
Conclusion:
The Fluent Bit vulnerabilities enable attackers to hide activity, corrupt evidence and even gain direct control of cloud workloads.
This puts cloud systems at risk because security teams may not see the real activity happening inside their environment.
Organizations using Fluent Bit should patch immediately, restrict network access and enforcing strong authentication and least‑privilege deployment as urgent priorities to reduce the risk of remote takeover and systemic observability compromise.
References:
Summary : Mozilla released the Firefox 145 Security Update on November 11, 2025, addressing 16 vulnerabilities affecting multiple components allowing arbitrary code execution.
The Mozilla Firefox advisory details reveal that exploiting these vulnerabilities requires attackers to deliver malicious content via compromised websites or through network attacks. The vulnerability landscape reveals concerning patterns in critical components where WebGPU graphics processing emerges as a significant attack surface, with five separate boundary condition flaws identified.
| OEM | Mozilla |
| Severity | High |
| Date of Announcement | 2025-11-11 |
| Actively Exploited | No |
| Exploited in Wild | No |
| Advisory Version | 1.0 |
Overview
| Vulnerability Name | CVE ID | Product Affected | Severity |
| Graphics Race Condition Enabling Remote Code Execution | CVE-2025-13012 | Firefox | High |
| WebAssembly Boundary Error RCE | CVE-2025-13016 | Firefox | High |
| WebGPU Boundary Error Leading to Remote Code Execution | CVE-2025-13021 | Firefox | High |
| WebGPU Boundary Error Leading to Remote Code Execution | CVE-2025-13022 | Firefox | High |
| WebGPU Sandbox Escape via Boundary Flaw | CVE-2025-13023 | Firefox | High |
| JavaScript Engine JIT Miscompilation RCE | CVE-2025-13024 | Firefox | High |
| WebGPU Boundary Error Leading to Remote Code Execution | CVE-2025-13025 | Firefox | High |
| WebGPU Sandbox Escape and Code Execution | CVE-2025-13026 | Firefox | High |
| Memory Safety Bugs Allowing Arbitrary Code Execution | CVE-2025-13027 | Firefox, Thunderbird | High |
Mozilla released the Firefox 145 Security Update on November 11, 2025, addressing 16 vulnerabilities affecting multiple components.
The patched vulnerabilities include memory safety bugs, boundary errors, race conditions, and sandbox escapes that could allow remote code execution (RCE), privilege escalation or data exposure. Although no active exploitation has been reported. Users and administrators should upgrade immediately to prevent exploitation and maintain browser security integrity.
Technical Summary
Memory corruption and sandbox escape issues could allow attackers to run malicious code or bypass.
Firefox’s isolation controls, leading to full system compromise. WebRTC and multimedia use-after-free bugs further increase the risk of crashing or leaking live data.
Though no exploitation has been detected, users and enterprises should update immediately to reduce exposure to emerging browser-based threats.
| CVE ID | System Affected | Vulnerability Details | Impact |
| CVE-2025-13012 | Graphics Subsystem | Race condition vulnerability leads to system crash or remote execution. | Remote Code Execution |
| CVE-2025-13016 | JavaScript: WebAssembly | Incorrect memory boundary validation allows code execution through crafted scripts. | Remote Code Execution |
| CVE-2025-13021 | Graphics: WebGPU | Boundary condition errors may cause memory corruption and remote code execution. | Remote Code Execution |
| CVE-2025-13022 | Graphics: WebGPU | Boundary error in WebGPU rendering pipeline allowing memory corruption | Remote Code Execution |
| CVE-2025-13023 | Graphics: WebGPU | Sandbox escape via boundary flaw enabling code execution outside browser process. | Sandbox Escape, RCE |
| CVE-2025-13024 | JavaScript Engine: JIT | JIT miscompilation allows stealthy remote code execution within browser context. | Remote Code Execution |
| CVE-2025-13025 | Graphics: WebGPU | Improper memory boundary control in WebGPU leads to RCE. | Remote Code Execution |
| CVE-2025-13026 | Graphics: WebGPU | Sandbox escape enabling remote code execution beyond browser sandbox. | Sandbox Escape, RCE |
| CVE-2025-13027 | Firefox / Thunderbird | Memory safety errors across multiple components allow arbitrary code execution. | Remote Code Execution |
Source: Microsoft, Cybersecurity News
In addition to several high severity vulnerabilities, the update also addresses several other Medium and Low severity vulnerabilities across browser subsystems –
Recommendations:
Here are some recommendations below
Conclusion:
The Firefox 145 update is a critical security release addressing sixteen vulnerabilities across graphics, WebGPU, JavaScript engine, DOM, and WebRTC components.
Immediate patching and adherence to strong security hygiene are essential to prevent remote code execution, sandbox escapes, and data leaks. Timely remediation ensures operational continuity and protection of both individual and enterprise users against evolving exploitation techniques targeting browser flaws.
Improving browsing behavior significantly reduces risk exposure of users. Reporters Oskar L and Jamie Nicol highlighted how these bugs exploit WebGPU’s high-performance rendering, a feature increasingly targeted as web apps grow more graphics-intensive.
References:
Summary : Microsoft’s November 2025 Patch Tuesday resolves 63 vulnerabilities across multiple Microsoft components. The Microsoft Patch Tuesday also addresses four “Critical” vulnerabilities, two of which are remote code execution vulnerabilities, one is an elevation of privileges and the fourth is an information disclosure flaw.
| OEM | Microsoft |
| Severity | Critical |
| Date of Announcement | 2025-11-11 |
| No. of Patches | 63 |
| Actively Exploited | Yes |
| Exploited in Wild | Yes |
| Advisory Version | 1.0 |
Overview : Key Updates on Patch Tuesday
The update includes one actively exploited zero-day vulnerability (CVE-2025-62215) in the Windows Kernel and five additional Critical-rated vulnerabilities affecting Office, DirectX, GDI+, Visual Studio, and Nuance PowerScribe.
This release continues Microsoft’s focus on privilege escalation and remote code execution (RCE) vulnerabilities, highlighting the urgent need for comprehensive patch management across enterprise systems.
Here are the CVE addresses for Microsoft & non-Microsoft:
Breakdown of October 2025 Vulnerabilities
Source: Microsoft
| Vulnerability Name | CVE ID | Product Affected | Severity | CVSS Score |
| Windows Kernel Elevation of Privilege Vulnerability (Zero-Day, Exploited in Wild) | CVE-2025-62215 | Windows 10, 11, Server 2016–2022 | Critical | 9.0 |
| Microsoft Office Use-After-Free Remote Code Execution Vulnerability | CVE-2025- 62199 | Microsoft Office (Word/Excel/Office Suite) | Critical | 9.8 |
| Nuance PowerScribe Missing Authorization Information Disclosure Vulnerability | CVE-2025-30398 | Nuance PowerScribe 360 | Critical | 9.1 |
| Windows DirectX Graphics Kernel Use-After-Free Vulnerability | CVE-2025-60716 | Windows DirectX Graphics Kernel | Critical | 8.8 |
| Microsoft GDI+ Heap-Based Buffer Overflow RCE Vulnerability | CVE-2025-60724 | Microsoft Graphics Component (GDI+) | Critical | 8.7 |
| Visual Studio Command Injection Remote Code Execution Vulnerability | CVE-2025-62214 | Microsoft Visual Studio / Visual Studio Code | Critical | 8.1 |
Technical Summary
The zero-day is a Windows Kernel bug that lets attackers gain full system control. Other critical & important vulnerabilities include Office and GDI+ vulnerabilities that could allow hackers to run malicious code or steal data.
Microsoft also patched issues in Visual Studio, DirectX, and Azure services. Users and admins are strongly advised to install these updates right away to stay protected.
| CVE ID | System Affected | Vulnerability Details | Impact |
| CVE-2025-62215 | Windows Kernel | Race conditions in shared resource execution enables local attackers to elevate privileges to SYSTEM (Zero-Day; Exploited in Wild) | Elevation of Privilege |
| CVE-2025-62199 | Microsoft Office | Use-after-free vulnerability in Office allows RCE via malicious documents, typically delivered through phishing campaigns | Remote Code Execution |
| CVE-2025-30398 | Nuance PowerScribe 360 | Missing authorization vulnerability allows disclosure of sensitive medical or user data over the network | Information Disclosure |
| CVE-2025-60716 | Windows DirectX Graphics Kernel | Use-after-free conditions allow local attackers to escalate privileges, potentially compromising the entire system | Elevation of Privilege |
| CVE-2025-60724 | Microsoft GDI+ | Heap-based buffer overflow allows attackers to execute arbitrary code remotely via crafted network traffic or malicious files | Remote Code Execution |
| CVE-2025-62214 | Visual Studio | Command injection vulnerability allows attackers to execute arbitrary code locally in developer environments | Remote Code Execution |
Source: Microsoft
In addition to several other Important severity vulnerabilities were addressed below –
Source: Microsoft, bleepingcompute, cybersecuritynews
Key Affected Products and Services
The November 2025 security updates address critical and important vulnerabilities across a broad range of Microsoft products and services:
Updates for Kernel, Hyper-V, Kerberos, RRAS, WinSock, Smart Card, Bluetooth subsystems.
Patches for Word, Excel, and related components impacted by RCE and Information Disclosure vulnerabilities.
Fixes for Azure Monitor Agent, Dynamics 365, Entra ID, and related connectors.
Patches for GDI+, DirectX, WSL GUI.
Updates for Visual Studio, Visual Studio Code, and GitHub Copilot.
Patches for Nuance PowerScribe (Medical domain).
Updates for Microsoft OneDrive for Android.
Remediation:
Here are some recommendations below
Conclusion:
Microsoft’s November 2025 Patch Tuesday resolves 63 vulnerabilities, including one actively exploited Zero-Day and multiple Critical RCE and EoP vulnerabilities in Office, Windows Kernel, GDI+, and Visual Studio.
Given the confirmed exploitation and the presence of memory corruption vulnerabilities, immediate patch deployment is necessary to prevent potential ransomware and privilege escalation attacks in our modern cyber world.
References:
Summary : A critical unauthenticated access vulnerability in Triofox is being actively exploited in the wild by threat actor UNC6485. Attackers exploit a Host header spoofing vulnerability to bypass authentication, create native admin accounts and chain abuse of the built-in antivirus feature to execute arbitrary code under SYSTEM privileges.
| OEM | Gladinet |
| Severity | Critical |
| CVSS Score | 9.1 |
| CVEs | CVE-2025-12480 |
| POC Available | YES |
| Actively Exploited | YES |
| Exploited in Wild | YES |
| Advisory Version | 1.0 |
Overview
Triofox is an enterprise file-sharing and remote access platform by Gladinet that enables secure file sync, sharing, and collaboration across on-premises and cloud environments. Immediate upgrade is mandatory to prevent full system compromise, ransomware and persistent remote access.
| Vulnerability Name | CVE ID | Product Affected | Severity | Fixed Version |
| Unauthenticated Access via Host Header Spoofing & Antivirus RCE Chain | CVE-2025-12480 | Triofox | Critical | v16.7.10368.56560 or later |
Technical Summary
The vulnerability in the CanRunCriticalPage() function within GladPageUILib.dll, which allows access to setup pages, if the Host header is “localhost” – without validating the request origin. Attackers spoof this header externally to initiate the setup process, create a Cluster Admin account, and gain authenticated access.
Once logged in, attackers exploit the antivirus configuration feature, which allows arbitrary executable paths. By uploading a malicious script to a shared folder and setting it as the antivirus scanner, the file executes with SYSTEM-level privileges inherited from the Triofox service.
| CVE ID | System Affected | Vulnerability Details | Impact |
| CVE-2025- 12480 | Triofox < 16.7.10368.56560 | Host header attack bypasses authentication to AdminDatabase.aspx that enables admin account creation. Chained with antivirus path abuse to run uploaded payloads as SYSTEM | Authentication Bypass, Admin Account Creation, Remote Code Execution, Full System Compromise, Persistent Access, Data Exfiltration, Lateral Movement |
Indicators of Compromise (IOCs)
Host-Based Artifacts
| Artifact | Description | SHA-256 Hash |
| C:\Windows\appcompat\SAgentInst aller_16.7.10368.56560.exe | Installer containing Zoho UEMS Agent | 43c455274d41e58132be7f66139566a941190ceba46082eb 2ad7a6a261bfd63f |
| C:\Windows\temp\sihosts.exe | Plink | 50479953865b30775056441b10fdcb984126ba4f98af4f647 56902a807b453e7 |
| C:\Windows\temp\silcon.exe | PuTTy | 16cbe40fb24ce2d422afddb5a90a5801ced32ef52c22c2fc7 7b25a90837f28ad |
| C:\Windows\temp\file.exe | AnyDesk | ac7f226bdf1c6750afa6a03da2b483eee2ef02cd9c2d6af71e a7c6a9a4eace2f |
| C:\triofox\centre_report.bat | Attacker batch script filename | N/A |
Network-Based Artifacts
| IP Address | ASN | Description |
| 85.239.63[.]37 | AS62240 – Clouvider Limited | IP address of the attacker used to initially exploit CVE-2025-12480 to create the admin account and gain access to the Triofox instance |
| 65.109.204[.]197 | AS24950 – Hetzner Online GmbH | After a dormant period, the threat actor used this IP address to login back into the Triofox instance and carry out subsequent activities |
| 84.200.80[.]252 | AS214036 – Ultahost, Inc. | IP address hosting the installer for the Zoho UEMSAgent remote access tool |
| 216.107.136[.]46 | AS396356 – LATITUDE-SH | Plink C2 |
Source: cloud.google.com
Recommendations:
Upgrade Triofox to version 16.7.10368.56560 or latest from the official Gladinet portal.
Conclusion:
This vulnerability represents a severe supply-chain risk in enterprise file-sharing platforms, enabling zero-authentication RCE through misconfigured access controls and feature abuse. With active in-the-wild exploitation by UNC6485 and rapid post-patch attacks, delayed patching significantly increases breach likelihood.
Immediate upgrade, log monitoring, and network hardening are essential to prevent ransomware deployment, data theft, and network pivoting. This incident reinforces the need for secure-by-design input validation and principle of least privilege in remote access tools.
References:
Encryption is often taken as last line of defense and organizations are using encryption to secure their data. Understanding and adopting the latest encryption technologies is crucial for keeping data secure. In current scenario when attackers are equally lazed with latest technologies, companies can strengthen their cybersecurity strategies and continue to adapt encryption as last line of their defense. When organizations enhance their encryption practices today, they can protect their digital assets for the future.
As cyber attacks are evolving so as encryption advances. Now numerous key developments will shape the future of cybersecurity. Once inside the network, cyber criminals can easily view and steal sensitive data. If that data is encrypted, they have no way of accessing it without a decryption key, saving the data from being compromised.
For example, the continuous evolution of quantum computing presents challenges and opportunities for encryption. Quantum-resistant algorithms must increase in speed to enhance security against quantum attacks.
The FinWise Data Breach a Stark Example
On May 31, 2024, the ex-employee accessed FinWise Bank’s systems after leaving the company and leaked sensitive personal information belonging to 689,000 customers of American First Finance (AFF). Even more alarming, this unauthorized access went undetected for more than a year before being discovered by the bank on June 18, 2025.
The FinWise Data breach revealed lapses like time gap between the initial breach and its discovery. The Bank came to understand about the incident and notified affected customers in June 2025 which was over a year after the breach occurred. This was a huge time gap and lawsuits allege that the stolen data may not have been adequately encrypted and secured, causing public criticism and concern.
Security experts emphasize that a well-designed information protection framework must not only encrypt critical financial data but also proactively detect and prevent abnormal access attempts.
Quantum computing & Encryption
Organizations who relies on encryption to keep its critical business communications and data safe are secure now. But as per RAND, experts expect quantum computers capable of breaking today’s encryption standards to arrive by the 2030sOpens a new window .
In the latest updates The Federal Trade Commission (FTC) has sent letters to major tech companies in the United States, urging them to resist foreign governments’ demands to weaken encryption.
The letters were sent by FTC Chairman Andrew Ferguson to Akamai, Alphabet (Google), Amazon, Apple, Cloudflare, Discord, GoDaddy, Meta, Microsoft, Signal, Snap, Slack, and X.
Traditional encryption relies on math problems that would take classical computers centuries to solve. RSA encryption, which protects much of today’s internet traffic, works because factoring massive numbers is impossibly hard for regular computers. But tomorrow’s computers will make quick work of it. According to the MIT Technology Review, researchers have shown that a quantum computer with 20 million noisy qubits could crack RSA-2048 in just 8 hoursOpens a new window .
The question is Encryption alone is sufficient to protect data
As per researchers Encryption alone is no longer sufficient to protect privacy in LLM interactions, as metadata patterns can be exploited to infer sensitive subjects and corporate intent. Researchers at Microsoft have revealed a new side channel attack named Whisper Leak that can reveal the topic of encrypted conversations between users and language models, even without access to the underlying text.
The discovery highlights a growing blind spot in AI security where encryption alone no longer guarantees privacy in model interactions.
What we must know about Whisper Leak the side channel attack
Whisper Leak exploits often exploits a side channel in network communication rather than a flaw in encryption itself. LLM services generate responses step by step, by producing one token at a time instead of the entire response at once. Also, the communications with AI-powered chatbots are often encrypted with HPPS over TLS (HTTPS), ensuring the authenticity of the server and security through encryption.
A side channel attack breaks cryptography by using information leaked by cryptography, such as monitoring the electromagnetic field (EMF) radiation emitted by a computer screen to view information before it’s encrypted in a van Eck phreaking attack, aka Transient Electromagnetic Pulse Emanation STandard (TEMPEST).
Encryption the last line in defense & Helps Orgs Embrace GDPR
If sensitive information is no longer required, the best way to protect it is to delete it. However, when files are deleted from a hard drive they leave traces that can be reconstructed by thieves and hackers. By encrypting the files before deletion, the remnants that remain on the drive will remain encrypted and remain inaccessible should they be reconstructed. In this way, encryption protects your privacy, even when the files are gone.
Companies should, therefore, ensure that all devices leaving the workplace are encrypted. Most phones have a native encryption option that can be easily activated, while laptops can have either their hard drives or sensitive data encrypted depending on the tools an organization wants to use.
Nowadays data protection is no longer an option. Companies can’t ignore the problem and hope they won’t be targeted by malicious threat actors.
GDPR itself recommends encryption as an effective tool for data protection as do data protection standards such as the CIS Controls which advocate a data security strategy based on a combination of encryption, integrity protection and data loss prevention techniques.
At the end Encryption ensures that, whether these devices are lost, stolen or forgotten, the data on them is useless to anyone who tries to access it without a decryption key.
(Source: https://www.bleepingcomputer.com/news/security/finwise-data-breach-shows-why-encryption-is-your-last-defense/)
Sources: https://www.csoonline.com/
Summary: React Native is an open source framework maintained by Meta . A critical remote code execution vulnerability in the @react-native-community/cli package, a core toolset used by React Native developers. The flaw allows unauthenticated remote attackers to execute arbitrary OS commands on machines running the React Native Metro development server.
| Severity | Critical |
| CVSS Score | 9.8 |
| CVEs | CVE-2025-11953 |
| POC Available | Yes |
| Actively Exploited | No |
| Advisory Version | 1.0 |
Overview
A critical remote code execution vulnerability in the @react-native-community/cli package, a core toolset used by React Native developers. The flaw allows unauthenticated remote attackers to execute arbitrary OS commands on machines running the React Native Metro development server.
The vulnerability comes from unsafe input handling in the /open-url endpoint using the insecure open() function, and a React Native CLI flaw that exposes the server to remote code execution. Immediate updates and mitigations are recommended for all using the affected package versions.
| Vulnerability Name | CVE ID | Product Affected | Severity | Affected Version |
| OS Command Injection | CVE-2025-11953 | @react-native-community/cli @react-native-community/cli-server-api | Critical | @react-native-community/cli-server-api versions 4.8.0 through 20.0.0-alpha.2 |
Technical Summary
The Metro development server’s /open-url HTTP POST endpoint unsafely passes unsanitized user input (url field) as an argument to the open() function from the open NPM package which leads to OS command injection.
On Windows, the vulnerability allows arbitrary shell command execution with full control over parameters via cmd /c start command invocation. On macOS/Linux, arbitrary executables can be launched with limited parameter control. Further exploitation may lead to full RCE, but not confirmed yet. The server binds to all interfaces by default (0.0.0.0), exposing the endpoint externally to unauthenticated network attackers.
| CVE ID | Component Affected | Vulnerability Details | Impact |
| CVE-2025-11953 | Development Server’s /open-url Endpoint | The React Native CLI’s Metro server binds to external interfaces by default and exposes a command injection flaw, letting remote attackers send POST requests to run arbitrary executables or shell commands on Windows. | Remote OS Command Injection |
Recommendations
If upgrading is not possible,
How these kind of security flaw can cause damage?
This vulnerability poses a critical threat to React Native developers using the Metro development server due to unauthenticated RCE via network exposure. For any unauthenticated network attacker this is privilege they can weaponize the flaw and send a specially crafted POST request to the server. Then run arbitrary commands.
The attack takes a different turn when it comes to Windows and the exploitation is severe. The attackers can also execute arbitrary shell commands with fully controlled arguments, while on Linux and macOS, it can be widely used to execute arbitrary binaries with limited parameter control.
The vulnerable endpoint, /open-stack-frame, is designed to help developers open a file in their editor at a specific line number when debugging errors. This endpoint accepts POST requests with parameters such as file and lineNumber.
The incident highlight requirement for more rigorous input validation and secure-by-default configurations in developer environments.
What should organizations looks for while selecting a comprehensive tools that can provide thorough combing across their IT environment, networks, applications and cloud infrastructure.
Detecting vulnerabilities, misconfigurations with GaarudNode from Intruceptlabs makes it a go to scanner
References:
Overview : Brash Vulnerability works on Google Chrome and all web browsers that run on Chromium.
A newly disclosed vulnerability, Brash, exposed a critical architectural flaw in Chromium’s Blink rendering engine. Blink is Chromium’s open-source rendering engine responsible for parsing HTML, CSS, and JavaScript, building the DOM and render trees, and executing script-driven updates to the browser interface.
It underpins the user experience of all Chromium-based browsers and is a core component of their performance and stability.
The issue allows a malicious web page to crash Chromium-based browsers within seconds, including Chrome, Microsoft Edge, Brave, Opera etc. The attack works by overloading Blink’s main UI thread using a flood of unthrottled DOM operations. A public proof-of-concept (PoC) exploit is available and can be tested on machines, that escalating the urgency for patching across all Chromium-based platforms.
Technical Details
Blink lacks any rate limiting or coalescing on rapid document. title updates, allowing an attacker to flood the browser with millions of DOM mutations per second.
This saturates the browser’s main UI thread, causing extreme CPU usage and blocking event processing, which leads to the browser tab freezing or crashing within 15 to 60 seconds. The exploit can also be use to trigger after a delay or at a precise scheduled time, turning it into a highly controllable logic bomb.
The exploit requires no special permissions beyond navigating to a malicious page, presenting a severe and immediate operational risk until patches are deployed.
Attack Flow


Recommendations
You can follow the recommendations below
Conclusion:
The Brash vulnerability reveals how a simple architectural oversight. It lets attackers crash browsers by flooding them with too many title updates too fast, causing the browser to freeze or crash. This attack can be scheduled to happen later, making it harder to detect.
Mozilla Firefox and Apple Safari are immune to the attack, as are all third-party browsers on iOS, given that they are all based on WebKit.
The best defense is to keep browsers updated, avoid suspicious links and stay alert for unusual computer slowdowns.
References:
Docker Compose Path Traversal Vulnerability Enables Arbitrary File Write and System Compromise
Summary:
| OEM | Docker |
| Severity | High |
| CVSS Score | 8.9 |
| CVEs | CVE-2025-62725 |
| Date of Announcement | 2025-10-28 |
| Actively Exploited | No |
| Exploited in Wild | No |
| Advisory Version | 1.0 |
Overview
A high-severity path traversal vulnerability was identified in Docker Compose, a widely-used tool for defining and managing multi-container Docker applications.
This flaw occurs in the handling of remote OCI-based Compose artifacts, allowing an attacker to craft malicious artifact annotations that bypass directory restrictions. As a result, malicious files can be written outside the intended cache directory on the host system.
This vulnerability can be triggered even by seemingly harmless commands such as docker compose ps or docker compose config that resolve remote artifacts. Organizations should upgrade immediately to avoid possible system compromise.
| Vulnerability Name | CVE ID | Product Affected | Severity | CVSS Score |
| Path Traversal in OCI Artifacts Allowing Arbitrary File Write | CVE-2025-62725 | Docker Compose CLI | High | 8.9 |
Technical Summary
Docker Compose added support for fetching Compose files as OCI artifacts from remote registries. These artifacts contain layers with annotations indicating file paths for writing.
The vulnerability exists because Docker Compose did not sanitize or validate these path annotations prior to writing files, allowing path traversal sequences to escape the cache directory.
Attackers can exploit this by publishing malicious OCI artifacts with crafted annotations, leading to arbitrary file writes anywhere the Compose process has permissions, potentially overwriting sensitive files such as SSH authorized_keys, escalating privileges and compromising the host. The flaw affects Docker Compose versions prior to v2.40.2.
| CVE ID | System Affected | Vulnerability Details | Impact |
| CVE-2025-62725 | Docker Compose (Linux, Windows, macOS) | Path traversal via malicious remote OCI artifact annotations allowing arbitrary file write outside the Compose cache directory. | Arbitrary file write, potential system compromise, privilege escalation. |
Remediation
Apply security patches immediately to mitigate risks from privilege escalation and container escape.
Conclusion
Docker Compose vulnerability poses a serious risk of arbitrary file writes and system compromise through malicious OCI artifacts.
Due to the ease of exploitation when using remote Compose files, all users and organizations should upgrade to the patched Docker Compose version immediately, scrutinize remote artifact usage, and enhance their container security hygiene to mitigate this significant threat.
References
Summary
The CoPhish attack is a sophisticated phishing technique exploiting Microsoft Copilot Studio to steal OAuth tokens by tricking users into granting attackers unauthorized access to their Microsoft Entra ID accounts.
By Copilot Studio’s customizable AI agents, attackers create chatbots hosted on legitimate Microsoft domains that wrap traditional OAuth consent attacks in an authentic-looking interface, increasing the likelihood of successful deception.
Technical Details
The attackers often use a trial license or compromised tenant to create the agent, backdooring the authentication workflow so that, post-consent, OAuth tokens are exfiltrated via HTTP to attacker infrastructure.
Few Demo links like copilotstudio.microsoft.com add credibility, closely mimicking official Microsoft Copilot services, and victims see familiar branding and login flows.
While Microsoft has implemented consent policy updates including blocking risky permissions by default for most users significant gaps remain: unprivileged users can still approve internal apps and privileged admins retain broad consent authority.
Tokens exfiltrated by CoPhish can be used for impersonation, data theft or sending further phishing emails, often going undetected as the traffic is routed through Microsoft infrastructure.

malicious CopilotStudio page Source: securitylabs.datadoghq.com
Attack Flow
| Step | Description |
| 1. Build Malicious Copilot Agent | Attackers create a customized Copilot Studio chatbot, usually on a trial license within their own or a compromised Microsoft tenant, configuring it to appear as a legitimate assistant. |
| 2. Backdoor Authentication Workflow | The agent’s “Login” topic is modified to include an HTTP request that will exfiltrate any OAuth tokens granted by users during authentication. |
| 3. Share Demo Link | Attackers generate and distribute demo website URL (like, copilotstudio.microsoft.com) pointing to the malicious chatbot, mimicking official Copilot Studio services and passing basic domain trust checks. |
| 4. Victim and Trigger Consent | Victims access the link, interact with the familiar interface, and are prompted to login, beginning an OAuth consent flow that requests broad Microsoft Graph permissions. |
| 5. Token Exfiltration | After the victim consents, the agent collects the issued OAuth token and sends it via HTTP to an attacker-controlled server, often relaying through Microsoft IP addresses to avoid detection in standard traffic logs. |
| 6. Abuse Granted Permissions | Attackers use the stolen token to impersonate the victim, accessing emails, calendars, and files or conducting further malicious actions such as sending phishing emails or stealing sensitive data. |
| 7. Persist and Retarget | Due to policy gaps, attackers can repeat the process targeting both internal and privileged users, tailoring requested app permissions and adapting to Microsoft’s evolving security measures. |



Source: securitylabs.datadoghq.com
Why It’s Effective
Recommendations
Here are some recommendations below
Conclusion:
CoPhish highlights how AI-powered low-code platforms like Microsoft Copilot Studio can be exploited for advanced phishing attacks targeting identity systems.
Despite Microsoft’s improvements to consent policies, significant risks remain, requiring organizations to enforce strict consent controls, limit app creation, and monitor Entra ID logs vigilantly. As AI-driven tools grow, proactive security measures are essential to defend against these evolving hybrid threats leveraging trusted cloud services.
References:
Hashtags
#Infosec #CyberSecurity #Microsoft #Copilot #Vulnerabilitymanagement # Patch Management #ThreatIntel CISO #CXO #Intrucept
Recent Comments