🚨 CVE-2026-103630
Use after free in FedCM in Google Chrome prior to 154.0.8037.97 allowed a remote attacker to execute arbitrary code outside the sandbox via a crafted HTML page. (Chromium security severity: High)

🎖@cveNotify

🚨 CVE-2026-103631
Buffer overflow in WebRTC in Google Chrome prior to 154.0.8037.97 allowed a remote attacker to execute arbitrary code inside the sandbox via a crafted HTML page. (Chromium security severity: High)

🎖@cveNotify

🚨 CVE-2026-103648
Path traversal in image-downloader 4.3.0 allows an attacker who can control the download URL to cause downloaded response data to be written outside the configured destination directory.

🎖@cveNotify

🚨 CVE-2026-104843
uv is a Python package and project manager written in Rust. From 0.12.7 until 0.12.18, uv wheel extraction on Windows can process a malicious wheel in a way that writes a file outside the installation prefix, including an executable in a directory already present on the user's PATH. Non-Windows hosts are not affected. This issue is fixed in version 0.12.18.

🎖@cveNotify

🚨 CVE-2026-104844
PostCSS Selector Parser is a CSS selector parser that integrates with PostCSS but does not require it. Prior to 7.1.6, src/parser.js splitWord() can receive a flat selector as one word token carrying many class or ID indexes because period and hash characters are not tokenizer word delimiters. The uniqs() deduplication and per-index class and ID membership checks repeatedly scan the class and ID index arrays, while a separate Sass-interpolation filtering pass also performs repeated linear scanning. Together, these passes make parsing quadratic in the number of indexes and allow a crafted selector to occupy a synchronous parser thread. The maxNestingDepth guard does not mitigate the issue because the hostile selector can have zero nesting depth. Only consumers that synchronously parse untrusted selectors in an exposed request path are affected; ordinary build-time parsing of trusted sources is not affected. This issue is fixed in version 7.1.6.

🎖@cveNotify

🚨 CVE-2026-104845
Seroval facilitates JS value stringification, including complex structures beyond JSON.stringify capabilities. Prior to 1.6.3, deserializeTypedArray in fromJSON and fromCrossJSON trusts a deserialized source value as an ArrayBuffer and does not bound the serialized element count. An attacker can provide a small untrusted JSON object with a large length value, causing the array-like TypedArray constructor to synchronously allocate the selected number of elements and exhaust CPU or memory while starving the event loop. The offset check does not reject the crafted source because source.byteLength is undefined. DataView reaches a similar unchecked cast but throws rather than allocating, and the issue has no identified confidentiality or integrity impact. This issue is fixed in version 1.6.3.

🎖@cveNotify

🚨 CVE-2026-104846
Seroval facilitates JS value stringification, including complex structures beyond JSON.stringify capabilities. From 0.12.0 until 1.6.2, fromJSON deserialization of a fulfilled Promise control node can pass a plugin-produced callable-bearing thenable to a native Promise resolver. ECMAScript thenable assimilation then invokes the callable unexpectedly, allowing attacker-controlled JSON to trigger code in applications using plugin-capable Seroval releases. This path bypasses the Promise resolver type-confusion remediation in version 1.5.3 for CVE-2026-59940 because the unexpected invocation occurs through native Promise settlement after the referenced value is deserialized. This issue is fixed in version 1.6.2.

🎖@cveNotify

🚨 CVE-2026-104847
ProseMirror's view component renders and manages the editable browser interface for ProseMirror documents. Prior to 1.42.3, prosemirror-view paste handling accepts attacker-provided HTML whose clipboard slice context contains attributes that are not passed through schema attribute validation. When a user pastes the crafted HTML into an editor, the unvalidated context attributes can construct content that executes attacker-controlled JavaScript in the browser window containing the editor. This issue is fixed in version 1.42.3.

🎖@cveNotify

🚨 CVE-2026-104900
MISP contains a stored cross-site scripting (XSS) vulnerability in the index table rendering of the remote event preview. The count field template escaped the associated link URL but rendered the field value without HTML encoding. An attacker with the ability to create or modify events on a linked (remote) MISP server could craft an event identifier containing HTML or JavaScript markup. When a user on the local MISP instance views the remote event preview index, the unescaped value is rendered directly in the browser, allowing arbitrary script execution in the victim's session.

Preconditions:

- A linked/remote MISP server is configured and connected to the local instance.

- The attacker has sufficient access on the linked server to create or modify an event with a crafted identifier.

- A victim user on the local instance views the remote event preview index page.

Impact:

- Execution of arbitrary JavaScript in the context of the MISP web application.

- Potential session hijacking, data exfiltration, or unauthorized actions performed on behalf of the authenticated user.

Affected versions: <2.5.48.

🎖@cveNotify

🚨 CVE-2026-104901
MISP contains a cross-site scripting (XSS) vulnerability in the ID Translator feature. When a user views the ID Translator page, the application queries linked (remote) MISP servers for corresponding event identifiers. The event ID returned by the remote server was rendered in the HTML output without proper output encoding.

A malicious or compromised linked server could return a crafted event ID containing arbitrary HTML or JavaScript markup. This markup would be rendered in the browser of any user in the host organization who views the ID Translator page, enabling session hijacking, credential theft, or other client-side attacks.

Preconditions:

- The victim must be an authenticated user of the host MISP instance.

- A linked server must be configured on the host instance.

- The victim must navigate to the ID Translator page for a given event.

Affected versions: <2.5.48.

🎖@cveNotify

🚨 CVE-2026-104906
MISP contains a cross-site scripting (XSS) vulnerability in the TAXII object viewer. When displaying a remote TAXII object, the JSON content of string properties was rendered directly into an HTML pre block without HTML-encoding. An attacker who can control or influence the content of a TAXII object (e.g., by publishing a malicious object to a TAXII server that the victim's MISP instance subscribes to) can inject arbitrary HTML or JavaScript that executes in the context of the victim's MISP session.

Preconditions:

- The victim must be an authenticated MISP user with access to the TAXII object viewer.

- The victim must open or view the crafted TAXII object.

Impact:

- Execution of arbitrary JavaScript in the victim's browser within the MISP application context.

- Potential theft of session tokens, API keys, or other sensitive data accessible from the MISP interface.

- Potential for performing actions on behalf of the authenticated user.

Affected versions: <2.5.48.

🎖@cveNotify

🚨 CVE-2026-104907
MISP contains a cross-site scripting (XSS) vulnerability in the remote event preview page. When a linked (remote) MISP server is configured, the event preview renders tag identifiers inside an inline JavaScript onclick attribute. The tag ID value was HTML-escaped but not sanitized for the JavaScript string context, meaning a malicious linked server could supply a tag ID containing characters (such as a single quote) that break out of the JavaScript string literal and inject arbitrary script.

Preconditions:

- A linked/remote MISP server is configured and connected to the local instance.

- The linked server supplies a crafted tag ID in an event.

- An authenticated user views the event preview and interacts with the affected tag element.

Impact:

- Arbitrary JavaScript execution in the context of the viewing user's browser session, potentially allowing session hijacking, data exfiltration, or unauthorized actions on behalf of the user.

Affected versions: MISP prior to the fix commit (v2.5.48 or later, exact boundary unconfirmed).

🎖@cveNotify

🚨 CVE-2026-104908
MISP contains an improper input validation vulnerability in the decaying model import functionality. The import endpoint was intended to create a new decaying model belonging exclusively to the importing user's organisation, with the default flag forced to off.

However, the application stripped only the top-level id and uuid fields and pinned org_id and default on the outer array before saving the data flat. A user with decaying-model permissions could supply a nested model key carrying its own primary key, organisation identifier, and default flag, which bypassed those guards during the save operation.

Impact:

- A user with perm_decaying could overwrite an existing decaying model belonging to another organisation in place, altering its name, formula, parameters, or ownership.

- A user could create or modify a model flagged as the organisation default, affecting scoring behaviour for other users.

- A user could reassign a model's organisation to an arbitrary value.

Preconditions:

- Authenticated user with decaying-model permission (perm_decaying).

- Network access to the MISP instance.

Affected: <2.5.48.

🎖@cveNotify

🚨 CVE-2026-104910
MISP contains an authorization bypass in the related events listing functionality. When a user requests the list of events correlated to a given event, the system retrieved related event metadata directly from the correlation table without re-validating the caller's access rights against each related event.

The correlation table stores a snapshot of the event's distribution level and sharing group at the time the correlation was created, and does not carry the published flag. As a result, events that the caller is not permitted to open—because they are unpublished, or because their distribution or sharing group has changed since the correlation was recorded—were still returned with their metadata (title, date, correlating value counts).

Preconditions:

- An authenticated user with access to at least one event in MISP.

- The existence of correlation entries linking that event to other events the user should not be able to view.

Impact:

- Unauthorized disclosure of event metadata (titles, dates, correlation counts) for events the user has no right to access.

- Potential reconnaissance of threat-intelligence event names and timelines across sharing groups.

Affected: MISP versions prior to the fix commit (2ffa97f05).

🎖@cveNotify

🚨 CVE-2026-104914
MISP contains an improper access control vulnerability in its attribute search and paginated attribute view endpoints.

When a user queries for soft-deleted attributes (e.g., via the deleted-attributes search or the paginated attribute listing), the application returned soft-deleted attributes belonging to events owned by other organizations to any authenticated user who had visibility of the event. The event detail view correctly restricted soft-deleted attribute visibility to the owning organization and sync-permission users, but the attribute search and paginated view code paths lacked this restriction.

Preconditions:

- An authenticated MISP user with at least read access to an event owned by another organization.

- The user issues a query for deleted attributes (search or paginated view with the deleted filter).

Impact:

- Confidentiality: Soft-deleted threat intelligence attributes (e.g., IOCs, indicators, context) from other organizations are disclosed to unauthorized users. This may expose sensitive intelligence that the owning organization intended to remove from general visibility.

Affected versions: MISP versions prior to v2.5.48.

🎖@cveNotify

🚨 CVE-2026-51898
sinaptik-ai pandas-ai 3.0.0 is vulnerable to Code Injection in CodeExecutor.execute.

🎖@cveNotify

🚨 CVE-2026-51899
In SuperAGI v0.0.14 and prior, controller endpoints (/api/agents/create, /api/agents/schedule, /api/agents/delete, /api/agents/edit_schedule, /api/agents/stop_schedule) allow authenticated users from one organization to create, schedule, edit, stop, and delete agents belonging to a different organization's project. The endpoints accept a project_id parameter but do not verify that the project belongs to the authenticated user's organization.

🎖@cveNotify

🚨 CVE-2026-51901
SuperAGI up to 0.0.14 is vulnerable to Incorrect Access Control. The agent execution controller endpoint /api/agentexecutions/schedule allows authenticated users from one organization to schedule existing agents belonging to a different organization without proper authorization checks. The endpoint accepts an agent_id parameter but does not verify that the agent belongs to the authenticated user's organization.

🎖@cveNotify

🚨 CVE-2026-51904
SuperAGI up to v0.0.14 contains an improper access control vulnerability in the agent execution controller. In affected source snapshots, create_agent_execution and create_agent_run in superagi/controllers/agent_execution.py accept a caller-supplied agent_id and fail to verify that the referenced agent belongs to the authenticated user's organization. A remote authenticated attacker from one organization can create or start execution records for agents owned by another organization through /agentexecutions/add or /agentexecutions/add_run.

🎖@cveNotify

🚨 CVE-2026-51906
In TaskingAI v0.3.0 in the DALL-E 3 image generation tool save_url_image function, a path traversal vulnerability allows attackers to write downloaded images to arbitrary locations on the server filesystem by manipulating the project_id parameter.

🎖@cveNotify

🚨 CVE-2026-51907
In TaskingAI v0.3.0 in the QR Code Generator plugin save_base64_image function, a path traversal vulnerability allows attackers to write image files to arbitrary locations on the server filesystem by manipulating the project_id parameter.

🎖@cveNotify