CVE-2026-25517: 
Python vulnerability analysis and mitigation

CVE-2026-25517 is an improper permission handling vulnerability in Wagtail, an open-source Django-based CMS, affecting its admin preview endpoints. Due to a missing authorization check (CWE-862), an authenticated Wagtail admin user with knowledge of a model's fields can craft form submissions to obtain preview renderings of any page, snippet, or site setting object for which previews are enabled, using arbitrary data of their choosing. The vulnerability affects Wagtail versions before 6.3.6, 6.4 through 7.0.3, 7.1 through 7.1.2, and 7.2 through 7.2.1. It was published on February 3, 2026, and carries a CVSS v3.1 base score of 2.7 (Low) and a CVSS v4.0 base score of 5.1 (Medium) (GitHub Advisory).

The root cause is a missing authorization check (CWE-862) on Wagtail's preview endpoints (PreviewOnEdit and PreviewOnCreate views in wagtail/admin/views/generic/preview.py and wagtail/admin/views/pages/preview.py). Prior to the fix, these views did not inherit from PermissionCheckedMixin and did not verify whether the requesting user had change or add permissions for the target model before rendering a preview. An attacker with Wagtail admin access can craft a POST request to preview endpoints (e.g., wagtailadmin_pages:preview_on_add or wagtailadmin_pages:preview_on_edit) with arbitrary field data, causing the server to render the model's template with attacker-supplied content. The fix adds PermissionCheckedMixin to the preview view classes and enforces can_edit() / can_add_subpage() checks before processing preview requests (GitHub Advisory, Patch Commit).

Successful exploitation results in limited information disclosure: an attacker with Wagtail admin access can render previews of pages, snippets, or site settings using arbitrary data, potentially exposing database contents that would otherwise only be accessible to users with edit-level permissions over the affected model. The existing stored data of the targeted object itself is not directly exposed, but depending on the template logic, related database records may be revealed through the rendered preview output. There is no impact on integrity or availability, and exploitation is strictly limited to authenticated Wagtail admin users — ordinary site visitors cannot exploit this vulnerability (GitHub Advisory).

No public proof-of-concept exploit code is known to exist, and there is no evidence of active in-the-wild exploitation as of the time of publication (GitHub Advisory). The vulnerability is not listed in CISA's Known Exploited Vulnerabilities (KEV) catalog. The EPSS score is approximately 0.043%, reflecting a very low probability of exploitation in the near term. Exploitation requires authenticated access to the Wagtail admin interface and knowledge of the target model's field names, significantly limiting the attacker pool.

1. Gain Wagtail admin access: Obtain credentials for any user account with access to the Wagtail admin interface (e.g., a low-privileged editor account without edit rights over the target model).
2. Identify target model and fields: Enumerate the target model's field names through available documentation, source code review, or by inspecting other admin pages accessible to the attacker's account.
3. Identify the preview endpoint URL: Determine the preview endpoint for the target model, such as wagtailadmin_pages:preview_on_edit (e.g., /admin/pages/<page_id>/edit/preview/) or the snippet/site setting equivalent.
4. Craft a malicious form submission: Construct a POST request to the preview endpoint with arbitrary field values designed to trigger template rendering that may expose related database contents (e.g., populating a ForeignKey or related field reference that causes the template to query and display restricted data).
5. Submit the request and analyze the response: Send the crafted POST request using a browser, Burp Suite, or curl while authenticated. Inspect the rendered HTML preview response for sensitive database contents exposed through the template logic (GitHub Advisory).

• Logs: Unusual POST or GET requests to Wagtail preview endpoints (e.g., /admin/pages/<id>/edit/preview/, /admin/snippets/<app>/<model>/<id>/preview/, /admin/settings/<app>/<model>/preview/) from users who do not have edit permissions over the corresponding model; repeated preview requests with varying field data from the same authenticated session.
• Application Logs: Django request logs showing HTTP 200 responses to preview endpoints for users whose role should not grant edit access to the previewed model type.
• Network: Authenticated HTTP POST requests to preview endpoint paths with form-encoded body data containing field names matching restricted models, originating from unexpected user accounts.

Wagtail has released patched versions addressing this vulnerability: 6.3.6, 7.0.4, 7.1.3, 7.2.2, and 7.3. All affected deployments should upgrade to the appropriate patched version as soon as possible. No workaround is available — the vendor explicitly states that there is no configuration-based mitigation short of upgrading. As interim risk reduction, administrators should restrict Wagtail admin access to only trusted users and audit existing admin user permissions to minimize the number of accounts that could exploit this issue (GitHub Advisory).

The vulnerability was reported by researcher @thxtech and remediated by the Wagtail core team (gasman, RealOrangeOne, laymonage). The advisory was published at a Moderate severity level, reflecting the limited exploitability due to the admin-access prerequisite. No significant broader media coverage or notable community controversy has been observed, consistent with the low CVSS score and absence of active exploitation (GitHub Advisory).