| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| The MLX inference backend in Docker Model Runner on macOS uses the MLX-LM library, which unconditionally imports and executes arbitrary Python files from model directories via the model_file configuration field in config.json. When a model's config.json specifies a model_file pointing to a Python file, MLX-LM uses importlib to load and execute it with no trust_remote_code gate or equivalent safety check. The MLX backend runs without sandboxing, resulting in arbitrary code execution on the Docker host as the Docker Desktop user.
Any container on the Docker network can trigger this by calling the model-runner.docker.internal API to pull a malicious model from an attacker-controlled OCI registry and request inference. |
| Mantis Bug Tracker (MantisBT) is an open source issue tracker. In versions 2.28.1 and below, given any pre-existing XSS / HTML injection vulnerability, an attacker can bypass the Content Security Policy's script-src directive by uploading a crafted attachment to any issue that, when accessed via the file_download.php link, will be downloaded with a valid JavaScript MIME type resulting in script execution. The uploaded payload must be sniffed as a valid JavaScript MIME type by PHP finfo (see file_create_finfo() API function). Non-JavaScript MIME types will not get imported in a <script> tag by the browser, due to response header X-Content-Type-Options being set to nosniff, which requires all imported JavaScript files to be a valid JavaScript MIME type. This issue has been fixed in version 2.28.2. |
| Mantis Bug Tracker (MantisBT) is an open source issue tracker. In versions 2.11.0 through 2.28.1, a Stored XSS vulnerability is caused by incorrect escaping of a saved filter's owner, allowing an attacker to inject arbitrary HTML on systems where $g_show_user_realname = ON. Note that By default, only users with Manager access level or above can save their filters publicly. This issue has been fixed in version 2.28.2. If developers are unable to update immediately, they can work around this issue by preventing display of users' real names (set $g_ show_user_realname = OFF; in configuration), and restricting the ability to store filters (set $g_stored_query_create_threshold / $g_stored_query_create_shared_threshold to NOBODY). |
| BentoML is a Python library for building online serving systems optimized for AI apps and model inference. In versions 1.4.38 and prior, the build packaging workflow follows attacker-controlled symlinks inside the build context and copies the referenced file contents into the generated Bento artifact. If a victim builds an untrusted repository or other attacker-supplied build context, the attacker can place a symlink such as loot.txt -> /tmp/outside-marker.txt or a link to a more sensitive local file. When bentoml build runs, BentoML dereferences the symlink and packages the target file contents into the Bento. The leaked file can then propagate further through export, push, or containerization workflows. An attacker can exfiltrate local files from the build host into the Bento artifact, exposing secrets such as cloud credentials, SSH keys, API tokens, environment files, or other sensitive local configurations. Because Bento artifacts are commonly exported, uploaded, stored, or containerized after build, the leaked file contents can spread beyond the original build machine. This issue has been fixed in version 1.4.39. |
| JupyterHub is software that allows users to create a multi-user server for Jupyter notebooks. In versions 4.1.0 through 5.4.4, XSRF protection (updated in 4.1.0) inappropriately treated requests with Sec-Fetch-Mode: no-cors as same-origin requests, bypassing XSRF checks. The JSON API is not affected, only HTTP form endpoints, such as /hub/spawn and /hub/accept-share, meaning attackers could trigger server spawn (but not access the server) and if the attacker is a JupyterHub user permitted to share access to their server, cause a user to accept a share and have access to the attacker's server. This issue has been fixed in version 5.4.5. If developers are unable to immediately upgrade, they can temporarily mitigate this issue by dropping requests to JupyterHub with Sec-Fetch-Mode: no-cors if they are using a reverse proxy. |
| The vllm-metal inference backend in Docker Model Runner on macOS unconditionally sets trust_remote_code=True when loading model tokenizers, and runs without sandboxing. This causes transformers.AutoTokenizer.from_pretrained() to import and execute arbitrary Python files included in any model pulled from an OCI registry, resulting in arbitrary code execution on the Docker host as the Docker Desktop user when inference is triggered.
Any container on the Docker network can trigger this by calling the model-runner.docker.internal API to pull a malicious model and request inference. |
| An issue in ClipBucket v5 v.5.5.2 allows an attacker to execute arbitrary code via the Authentication interface, login page endpoint and HTTP response security headers components |
| Previously, CVE-2024-45337 fixed an authorization bypass for misused ssh server configurations; if any other type of callback is passed other than public key, then the source-address validation would be skipped. |
| For certain crafted inputs, a 'ed25519.PrivateKey' was created by casting malformed wire bytes, leading to a panic when used. |
| libheif is a HEIF and AVIF file format decoder and encoder. In versions 1.21.2 and prior, a crafted HEIF sequence file where the saiz box declares more samples than actually exist in the track's chunk table causes a heap-buffer-overflow (out-of-bounds read) in the SampleAuxInfoReader constructor. The SampleAuxInfoReader constructor iterates over saiz->get_num_samples() samples but doesn't validate that this count is consistent with the number of chunks in the chunks vector. When saiz declares more samples than the chunks cover, the loop increments current_chunk past chunks.size(), causing an out-of-bounds read on the chunks vector. The vulnerability is triggered during file parsing (heif_context_read_from_file) without any additional user interaction. Any application using libheif to open untrusted HEIF files is affected. This issue has been fixed in version 1.22.0. |
| libheif is a HEIF and AVIF file format decoder and encoder. In versions 1.21.2 and prior, a malformed HEIF sequence file can trigger an out-of-bounds read in core sequence parsing logic, causing DoS. A malformed file can have stco.entry_count == 0 (creating no chunks) while still passing validation because saio.entry_count == 0 matches, but with saiz.sample_count > 0 the SampleAuxInfoReader constructor still enters its loop. This leads to an out-of-bounds dereference on the empty chunks[0] in chunked mode. |
| An authentication logic vulnerability in multiple TP-Link range extenders allows an unauthenticated attacker on an adjacent network to manipulate a login parameter and reset the administrator password due to insufficient validation.
Successful exploitation allows an attacker to obtain full administrative control of the affected device, potentially impacting on confidentiality, integrity, and availability. |
| Homarr is an open-source dashboard. Prior to version 1.45.3, it was possible to craft an input which allowed privilege escalation and getting access to groups of other users due to missing sanitization of inputs in ldap search query. The vulnerability could impact all instances using ldap authentication where a malicious actor had access to a user account. Version 1.45.3 has a patch for the issue. |
| A bitwise shift vulnerability in Zephyr's PTP subsystem allows a remote attacker to cause undefined behavior and potential system crashes. An attacker sends a crafted PTP_MSG_MANAGEMENT message to set an unvalidated negative log_announce_interval value in the port's data set. When a subsequent PTP_MSG_ANNOUNCE message is processed, port_timer_set_timeout_random computes a timeout as NSEC_PER_SEC >> -log_seconds; if the attacker-supplied value is sufficiently negative (e.g., -127), the shift amount exceeds the 64-bit integer width, triggering undefined behavior in C. This can cause a system crash via a compiler-generated illegal instruction trap on some architectures, or produce an erroneous zero timeout leading to resource starvation loops or other logical errors. |
| Parsing arbitrary HTML which is then rendered using Render can result in an unexpected HTML tree. This can be leveraged to execute XSS attacks in applications that attempt to sanitize input HTML before rendering. |
| An issue was discovered in all versions of PCManFM-Qt starting from 1.1.0. When a regular file's path is passed as a URI in an org.freedesktop.FileManager1.ShowFolders D-Bus method call, PCManFM-Qt delegates to a different program (based on the file type) without user confirmation. This could be used to achieve code execution or circumvent network namespace restrictions. NOTE: those outcomes are potentially unwanted by most users; however, the behavior of the product does comply with the applicable specification, and a simplistic solution (ensuring that the URI does not name a regular file) may have adverse consequences for I/O. |
| authentik is an open-source identity provider. In versions prior to 2025.12.5 and 2026.2.0-rc1 through 2026.2.2, authenticated non-admin users with at least one OAuth2 access token can retrieve the client_secret of confidential OAuth2 providers they have previously authenticated against, exposing sensitive information to users without the correct permissions. This logic is GET /api/v3/oauth2/access_tokens/. The API response includes a nested provider object containing client_id and client_secret for providers configured with client_type: confidential, which should not be accessible to low-privilege users. This issue has been fixed in versions 2025.12.5 and 2026.2.3. |
| authentik is an open-source identity provider. In versions prior to 2025.12.5 and 2026.2.0-rc1 through 2026.2.2, the PATCH /api/v3/core/users/{pk}/ API allows a caller with change_user on a target user to assign arbitrary groups through UserSerializer, including groups with is_superuser=True, without requiring enable_group_superuser, leading to privilege escalation. This bypasses the stricter permission model enforced in group-management paths and enables delegated user-management permissions to escalate target users to administrator-equivalent privilege. Users with permissions to update groups or permissions to update users are able to add themselves or other users they have permissions on to users which have superuser permissions. This issue has been fixed in versions 22025.12.5 and 2026.2.3. |
| TypeBot is a chatbot builder tool. Versions 3.15.2 and prior contain a critical stored XSS vulnerability in the app.typebot.io profile picture upload form. The application fails to sanitize or restrict SVG/XML-based uploads and directly renders them when accessed through the domain. By uploading a crafted malicious SVG file containing embedded JavaScript, an attacker will execute arbitrary JavaScript code. This vulnerability directly enables stored XSS exploitation because the payload is persistently stored on your infrastructure (app.typebot.io) and accessible from a public-facing, permanent link. Stored XSS via malicious SVG uploads to app.typebot.io allows attackers to execute arbitrary JavaScript in victims' browsers, enabling session/token theft, account takeover, and exfiltration of sensitive user data. This issue has been fixed in version 3.16.0. |
| Cross Site Scripting vulnerability in Advantech WebAccess/SCADA 8.0-2015.08.16 allows a remote attacker to obtain sensitive information via the decryption field in the Create New Project User component |
