Juho Forsén

The software fails to invalidate the stored AD/LDAP bind password when other connection settings, such as the server address, are modified. This allows a System Console administrator to inadvertently disclose the plaintext bind password to an attacker-controlled server through the Test Connection / diagnostics feature.

An issue exists where the system fails to require the re-entry of the stored client secret during the validation of an outgoing OAuth connection. This allows an authenticated user with permissions to manage outgoing OAuth connections to obtain the secret by directing the token URL to an attacker-controlled endpoint and triggering the validation process.

The application fails to validate the Inter-Process Communication (IPC) sender within the `leaveCall` handler. This allows a compromised Mattermost server or a user with script access to a connected server view to disconnect an active call associated with a different connected server by sending a `desktopAPI.leaveCall` IPC message.

The application fails to validate the URL scheme when verifying if a target URL is internal to the connected server. This allows a network-positioned attacker to load a plugin popout window over an insecure connection by using a link with a downgraded URL scheme.

The application fails to redact the pre-auth secret during the generation of a diagnostics report. A local attacker with access to the user's log files or diagnostics report can obtain the plaintext pre-auth secret for a connected server by inspecting the Server Connectivity (Step-3) diagnostics output.

An issue exists where markdown image rendering restrictions are not properly applied to AI bot tool result posts. This allows an authenticated attacker to exfiltrate data to an external server by injecting markdown image syntax into the tool result content that is subsequently rendered by a victim's client.

The software fails to validate path parameters when constructing API route paths. This allows an attacker to redirect API calls to unintended endpoints by using crafted IDs that include path traversal components, which are sequences of characters used to access files or directories outside the intended folder.

An authenticated attacker can cause server memory exhaustion and denial of service by uploading a specially crafted 7zip file containing excessive folder declarations. This occurs because the software fails to validate the 7zip archive structure before processing it.

An issue exists where the software fails to validate the `Host` header when constructing response URLs for custom slash commands. This allows an authenticated attacker to redirect slash command responses to a server under their control by using a spoofed `Host` header.

SWIG file names containing 'cgo' and well-crafted payloads could lead to code smuggling and arbitrary code execution at build time due to trust layer bypass.

The software does not properly limit memory allocation when processing PSD image files. This allows an authenticated attacker to exhaust server memory and cause a denial of service by uploading a specially crafted PSD file.

The software does not properly validate user permissions when creating Jira issues from Mattermost posts. This allows an authenticated attacker with access to the Jira plugin to read post content and attachments from channels they should not have access to. The issue occurs when using the `/create-issue` API endpoint and providing the post ID of an inaccessible post. The vulnerable parameter is the `post ID`.

Mattermost does not properly limit memory allocation when processing DOC files. This allows an authenticated attacker to exhaust server memory and cause a denial of service by uploading a specially crafted DOC file.

Mattermost versions 10.11.0 through 10.11.9 do not properly validate channel membership when retrieving data, potentially allowing a deactivated user to learn team names they should not have access to. This occurs due to a race condition in the `/common teams` API endpoint.

Mattermost fails to verify that post actions invoking '/share-issue-publicly' were created by the Jira plugin. This allows a malicious Mattermost user to potentially exfiltrate Jira tickets when victim users interact with affected posts.

The software does not properly validate user channel membership when attaching Mattermost posts as comments to Jira issues. This allows an authenticated attacker who has access to the Jira plugin to read post content and attachments from channels they are not authorized to access.

The Mattermost Jira plugin does not properly enforce authentication and issue-key path restrictions. This allows an unauthenticated attacker, knowing a valid user ID, to send authenticated GET and POST requests to the Jira server using specially crafted plugin payloads. These payloads can spoof the user ID and inject arbitrary issue key paths. The vulnerability requires the Jira plugin to be enabled.

The Mattermost application does not properly enforce access permissions within the Agents plugin. This allows other users to determine when users have read channels by accessing channel member objects.

The Mattermost application does not properly validate the relationship between a post being updated and the Microsoft Teams (MSTeams) plugin OAuth flow. This allows an attacker to modify any post by using a specially crafted MSTeams plugin OAuth redirect URL. The issue involves a failure to validate the connection between the post update and the OAuth process.

The software does not properly limit the dictionary size when processing RAR files. An attacker can exploit this by providing a specially crafted RAR file with a large dictionary size, leading to a Denial of Service due to an Out Of Memory crash.