Currently browsing: Technology

Overview
Multiple vulnerabilities were discovered in The Librarian, an AI-powered personal assistant tool provided by the company TheLibrarian.io. The Librarian can be used to manage personal email, calendar, documents, and other information through external services, such as Gmail and Google Drive, and also summarize meetings and schedule emails.
The vulnerabilities are among internal tools that The Librarian uses during its normal functions.
These tools, view_document, web_fetch, and image_generate, allow an authenticated user
* access to the administrative console
* internal web crawling and port scanning of the internal infrastructure for thelibrarian.io
* disclosure of the internal system prompt for The Librarian
All vulnerabilities have since been fixed by thelibrarian.io, and the tools have now been deprecated.
Description
TheLibrarian.io is an AI company that offers the namesake AI-powered personal assistant tool, “The Librarian”. This assistant can perform a variety of services and can integrate with other external applications. Some of these abilities include calendar management, sending email, and document management. Integratable services include Google products such as Gmail and Google Drive.
A series of vulnerabilities have been discovered within The Librarian that enable an attacker to access the internal infrastructure of TheLibrarian.io, including the administrator console and cloud environment. They also permit disclosure of the internal system prompt, web crawling, log access, and viewing of internal processes that infrastructure for TheLibrarian.io is running.
Below is a list of all the vulnerabilities and respective CVE IDs assigned to them.
VU#383552.1
The Librarian image_generation tool can be used to disclose the full system prompt through requesting an image to be generated with the embedded system prompt.
VU#383552.2
The Librarian view_document tool can be used to disclose the full system prompt through requesting the system prompt be appended to a document that is uploaded to the system.
CVE-2026-0612
The Librarian contains a information leakage vulnerability through the web_fetch tool, which can be used to retrieve arbitrary external content provided by an attacker, which can be used to proxy requests through The Librarian infrastructure.
CVE-2026-0613
The Librarian contains an internal port scanning vulnerability, facilitated by the web_fetch tool, which can be used with SSRF-style behavior to perform GET requests to internal IP addresses and services, enabling scanning and metadata retrieval of the Hertzner cloud environment that TheLibrarian uses.
CVE-2026-0615
The Librarian supervisord status page can be retrieved by the web_fetch tool, which can be used to retrieve running processes within The Librarian backend.
CVE-2026-0616
The Librarian’s web_fetch tool can be used to retrieve the Adminer interface content, which can then be used to log in to the internal backend system of The Librarian.
Impact
An attacker who exploits these vulnerabilities could control a wide variety of aspects of the internal infrastructure for TheLibrarian.io. This could include process control, lateral movement, and credential theft. CVE-2026-0614, CVE-2026-0615, and CVE-2026-0616 are largely responsible for this potential impact. VU#383552.1 to VU#383552.4 allow for exploitation and potential misuse of the capabilities of The Librarian, and could result in jailbreaks or unintended actions by the AI.
Solution
The vendor has stopped the web-fetch tool from being able to retrieve dangerous content. Web-retrieval is now handled by a third-party service. The vendor also stated that: “prompt content is not a secrecy boundary in our threat model” in regard to system prompt disclosure.
Acknowledgements
Thanks to the reporter, Aaron Portnoy of Mindgard.ai. This document was written by Christopher Cullen.

Overview
A vulnerability, tracked as CVE-2025-14894, has been discovered within Livewire Filemanager, a tool designed for usage within Laravel applications. The Livewire Filemanager tool allows for users to upload various files, including PHP files, and host them within the Laravel application. When a user uploads a PHP file to the application, it can be accessed and executed by visiting the web-accessible file hosting directory. This enables an attacker to create a malicious PHP file, upload it to the application, then force the application to execute it, enabling unauthenticated arbitrary code execution on the host device.
Description
Livewire Filemanager is a tool designed to be embedded into Laravel applications, allowing for files to be uploaded, stored and managed. Laravel is a PHP framework, intended for web application development. A vulnerability has been discovered within the Livewire Filemanager that enables remote code execution (RCE) by uploading a malicious PHP file. This vulnerability is tracked as CVE-2025-14894, and its description is below:

Livewire Filemanager, commonly used in Laravel applications, contains LivewireFilemanagerComponent.php, which does not perform file type and MIME validation, allowing for RCE through upload of a malicious php file that can then be executed via the /storage/ URL if a commonly performed setup process within Laravel applications has been completed.

As a note, Livewire Filemanager defines validation of file types to be out of scope, and recommends users perform their own file type validation. However, the ability to remotely execute files uploaded through the web application is what actually enables executing the malware.
During default usage of Livewire Filemanager, files can be accessed via the publicly accessible “storage/app/public” URL. This occurs if the php artisan storage:link command has previously been executed, enabling web serving. If a malicious PHP file is uploaded to the file manager, it can then be accessed and executed from that URL when passed a user ID alongside the request, enabling remote code execution on the target device.
Impact
The vulnerability enables unauthenticated remote code execution as the web server user, enabling full read and write of files accessible to that user, as well as the capability to further pivot and compromise connecting devices, making CVE-2025-14894 a high impact vulnerability.
Solution
At the time of writing, the vendor has not acknowledged the vulnerability. CERT/CC recommends using increased caution with Laravel Filemanager, and to check if the php artisan storage:link command has previously been executed, and if so, consider removing the web serving capability of the tool.
Acknowledgements
Thanks to the reporter HackingByDoing. This document was written by Christopher Cullen.

Overview
Redmi Buds, a series of Bluetooth earbuds produced and sold by Xiaomi, contain an Information Leak vulnerability and a Denial of Service (DoS) vulnerability in versions 3 Pro through 6 Pro. An attacker within Bluetooth radio range can send specially crafted RFCOMM protocol interactions to the device’s internal channels without prior pairing or authentication, enabling the exposure of sensitive call-related data or triggering repeatable firmware crashes.
Description
The two vulnerabilities originate from the firmware’s improper management of RFCOMM control and signaling mechanisms. The product specifications for Redmi Buds advertise support for the Bluetooth Classic profiles HFP, A2DP, and AVRCP, but direct experimentation has also confirmed the presence of additional, undocumented L2CAP/ RFCOMM channels active on the devices. These channels were likely implemented to facilitate auxiliary services or legacy audio support; vendor-specific internal interfaces are not required to be publicized in consumer product pages.
CVE-2025-13834 This vulnerability is the result of flawed bounds checking during the internal handling of abnormal RFCOMM TEST commands. When the device’s control channel (DLCI 0) receives a TEST command with a large length field but an empty payload, its faulty response handler returns a buffer of uninitialized memory. An attacker can exploit this behavior to steal up to 127 bytes of potentially sensitive data, such as the phone number of a user’s active call peer, with a single packet. Notably, the mechanism of this vulnerability is closely related to the infamous Heartbleed bug (CVE-2014-0160). Like Heartbleed, this flaw originates from blind trust in a packet’s length field without adequate bounds checking, resulting in an out-of-bounds read and unintended memory disclosure.
CVE-2025-13328 This vulnerability is caused by the firmware’s susceptibility to flooding attacks over RFCOMM channels. When an attacker floods the standard control channel (DLCI 0) with a high volume of legitimate TEST commands, the device’s processing queue is overwhelmed, leading to resource exhaustion and a firmware crash that forcibly terminates paired user connections. Other active data channels across the device’s RFCOMM implementation are also vulnerable to flooding via MSC (Modem Status Command) signaling frames, including both the standard HFP (Hands-Free Profile) channel and an undocumented Airoha auxiliary service channel.
Impact
Both vulnerabilities can be exploited by an unpaired, unauthenticated attacker within Bluetooth radio range without prior user interaction. The only prerequisite for exploitation is obtaining the MAC address of the target device, which can be discovered through basic Bluetooth sniffing tools. During testing with standard dongles and no additional signal amplification, exploitation was achieved at an approximate distance of twenty meters; however, physical barriers and Bluetooth version differences can be expected to influence the effective range.
The uninitialized memory dumped by CVE-2025-13834 threatens the confidentiality of Redmi Buds users during or after private calls, as demonstrated by a PoC used to obtain the phone number of a user’s active call peer. Any other metadata that utilizes this memory pool is vulnerable to exposure, and the attack can be triggered repeatedly without alerting the user. This vulnerability demonstrates the susceptibility of IoT protocol stacks to the class of “missing bounds check” or “buffer over-read” flaws famously exemplified by Heartbleed in web servers. It represents a critical oversight in memory management within the Redmi Buds Bluetooth firmware.
The DoS vulnerability, CVE-2025-13328, can be exploited to the detriment of device availability for legitimate users, inducing repeatable firmware crashes that forcibly disconnect all paired devices. To restore functionality, the earbuds must be physically reset by returning them to the charging case.
Solution
Xiaomi could not be reached for statements regarding remediation plans or mitigation guidance. To reduce exposure, users are advised to disable Bluetooth when the earbuds are not in use, particularly in public or shared environments.
Acknowledgements
Thanks to Choongin Lee, Jiwoong Ryu, and Heejo Lee for discovering, researching, and reporting these vulnerabilities. This document was written by Molly Jaconski.

Overview
The BeeS Examination Tool (BET) portal from BeeS Software Solutions contains an SQL injection vulnerability in its website login functionality. More than 100 universities use the BET portal for test administration and other academic tasks. The vulnerability enables arbitrary SQL commands to be executed on the back-end database, making an attacker able to manipulate the database, extract sensitive student data, and further compromise the host infrastructure. BeeS Software Solutions has since remediated the vulnerability, and no actions are necessary for customers at this time.
Description
Numerous universities implement the BET portal to unify the various tasks associated with administering examinations to students. Each university maintains their own instance of the BET portal, receiving updates from BeeS Software Solutions.
A vulnerability, tracked as CVE-2025-14598, was discovered within the login functionality of the portal. This vulnerability, facilitated by insufficient user input validation, enables arbitrary SQL injection. When exploited, an attacker can manipulate the backend database, steal student data (including credentials), and perform lateral movement, further compromising the host infrastructure.
BeeS Software Solutions issued a patch to all instances using the BET portal, changing code, enabling input validation, and changing various security settings to prevent exploitation and unauthorized access. All BET clients automatically received these changes.
Impact
The vulnerability permits an unauthenticated, remote attacker to achieve various results, including unauthorized database access, credential theft, potential lateral movement into infrastructure, acquisition of sensitive student and institutional data, and system-level access to the affected server.
Solution
No actions are needed by clients, as configurations and updated dynamic link libraries (DLLs) have been automatically installed and updated through ePortal : Secure Build (October 2025). Testing indicates that the changes successfully mitigated the vulnerability.
Acknowledgements
Thanks to the reporter, Mohammed Afnaan Ahmed, for reporting these vulnerabilities. This document was written by Christopher Cullen.

Overview
A flaw in the firmware-upload error-handling logic of the TOTOLINK EX200 extender can cause the device to unintentionally start an unauthenticated root-level telnet service. This condition may allow a remote authenticated attacker to gain full system access.
Description
In the End-of-Life (EoL) TOTOLINK EX200 firmware, the firmware-upload handler enters an abnormal error state when processing certain malformed firmware files. When this occurs, the device launches a telnet service running with root privileges and does not require authentication. Because the telnet interface is normally disabled and not intended to be exposed, this behavior creates an unintended remote administration interface.
To exploit this vulnerability, an attacker must already be authenticated to the web management interface to access the firmware-upload functionality. Once the error condition is triggered, the resulting unauthenticated telnet service provides full control of the device.
CVE-2025-65606
An authenticated attacker can trigger an error condition in the firmware-upload handler that causes the device to start an unauthenticated root telnet service, granting full system access.
Impact
A remote authenticated attacker may be able to activate a root telnet service and subsequently take complete control of the device. This may lead to configuration manipulation, arbitrary command execution, or establishing a persistent foothold on the network.
Solution
TOTOLINK has not released an update addressing this issue, and the product is no longer maintained. Users should restrict administrative access to trusted networks, prevent untrusted users from accessing the management interface, monitor for unexpected telnet activity, and plan to replace the vulnerable device.
Acknowledgements
Thanks to the reporter Leandro Kogan for bringing this to our attention. This document was written by Timur Snoke.

Overview
A vulnerability in the Forcepoint One DLP Client allows bypass of the vendor-implemented Python restrictions designed to prevent arbitrary code execution. By reconstructing the ctypes FFI environment and applying a version-header patch to the ctypes.pyd module, an attacker can restore ctypes functionality within the bundled Python 2.5.4 runtime, enabling direct invocation of DLLs, memory manipulation, and execution of arbitrary code.
Description
The Forcepoint One DLP Client (version 23.04.5642 and potentially subsequent versions) shipped with a constrained Python 2.5.4 runtime that omitted the ctypes foreign function interface (FFI) library. Although this limitation appeared intended to mitigate malicious use, it was demonstrated that the restriction could be bypassed by transferring compiled ctypes dependencies from another system and applying a version-header patch to the ctypes.pyd module. Once patched and correctly positioned on the search path, the previously restrained Python environment would successfully load ctypes, permitting execution of arbitrary shellcode or DLL-based payloads.
Forcepoint acknowledged the issue and indicated that a fix would be included in an upcoming release. According to the Forcepoint’s published knowledge base article (KB 000042256), the vulnerable Python runtime has been removed from Forcepoint One Endpoint (F1E) builds after version 23.11 associated with Forcepoint DLP v10.2.
Impact
Arbitrary code execution within the DLP client may allow an attacker to interfere with or bypass data loss prevention enforcement, alter client behavior, or disable security monitoring functions. Because the client operates as a security control on enterprise endpoints, exploitation may reduce the effectiveness of DLP protections and weaken overall system security.
The complete scope of impact in enterprise environments has not been fully determined.
Solution
Forcepoint reports that the vulnerable Python runtime has been removed in Endpoint builds after version 23.11 (Forcepoint DLP v10.2).
Users should upgrade to Endpoint versions which have been validated to no longer contain python.exe.
Acknowledgements
Thanks to the reporter, Keith Lee.
This document was written by Timur Snoke.