CVE-2026-95698

MISP registers PHP's phar stream wrapper in both its web and console entry points. This causes filesystem operations on caller-influenced paths to trigger implicit unserialize() (a deserialization sink), and a relocated application root can reach executable code inside an uploaded phar file, enabling code execution as the web user.

A typo in the MISP ACLComponent access control configuration caused the ACL rule for the previewEventAttributes action to reference the permission string 'theming_enabled*' (with a trailing asterisk) instead of the correct 'theming_enabled'. The malformed key does not match any valid permission identifier, causing the access control check for previewEventAttributes to malfunction — potentially resulting in either an authorization bypass or denial of access for legitimate users.

In MISP's UsersController login(), the pre-auth query for TOTP verification did not include the User.disabled column, so the disabled flag was not checked in that branch. Although later identify() validation still rejects disabled accounts, a TOTP-enabled user can proceed to the verification step.

In MISP, the __uploadLogo method in OrganisationsController processes a caller-supplied tmp_name with filesystem probes before verifying it is a real upload. An authenticated site-admin can supply an arbitrary server file path, allowing to check for file existence and determine image types.

In MISP, the __statisticsOrgs method uses the organization name directly as a file path component when checking for logo existence. Although the target directory does not exist, the pattern is a path traversal weakness that could allow probing arbitrary files if the directory existed.

MISP contains an authorization flaw in the Organisation model's captureOrg method. When the $force parameter is set to true, the method unconditionally overwrites organization metadata fields without verifying that the invoking user holds sufficient privileges. A user with a sharing group (SG) editor role can trigger this code path, allowing them to modify organization metadata that should be restricted to site administrators or users with sync permissions.

In MISP, the EventReport::uploadPicture method processes a caller-supplied tmp_name field by invoking file_exists(), mime_content_type(), and exif_imagetype() on the supplied path before verifying that the value was a genuine PHP upload via is_uploaded_file(). An authenticated user holding the perm_add permission could supply an arbitrary filesystem path as the tmp_name value. The application would then probe that path and return distinct validation error messages depending on whether the file existed, its MIME type, or its image format, allowing enumeration of file existence and type.

MISP contains an access control flaw in the EventReports functionality. The replaceSuggestionInReport action, which allows modification of suggestion content within an event report, was incorrectly mapped to the wildcard permission ('*') in the ACLComponent, making it accessible to any authenticated user regardless of their assigned permissions. All analogous report-modification actions correctly required the perm_add permission, while read-only actions such as downloadAsPDF appropriately used the wildcard.

In MISP, the Overmind event view enriches an event with its most recent attached report for preview purposes. The enrichment logic fetched the report using only the event ID as the lookup condition, without applying the report's own distribution/ACL constraints. Because MISP reports carry an independent distribution setting that can be more restrictive than the parent event's distribution, an authenticated user who had read access to the event could retrieve and view a report whose distribution explicitly excluded that user. Additionally, the query did not filter out soft-deleted reports, allowing retrieval of reports that had been administratively removed.

MISP contains a stored cross-site scripting (XSS) vulnerability in the admin email composition screen. The MISP.org organization name is interpolated directly into a JavaScript string literal without proper output encoding, allowing injection of arbitrary JavaScript. The injected script executes in the context of any authenticated user who loads the page, potentially enabling session hijacking, data exfiltration, or privileged actions.