Wagtail 8.0.1 release notes

October 20, 2026

What’s new

Security fix: Cross-Collection authorization bypass via Collection privacy settings

An authenticated user with access to the Wagtail administrative interface and the “change” permission to at least one Collection could change the privacy settings for a Collection they do not have access to, leading to them potentially removing access restrictions on sensitive documents. The vulnerability is not exploitable by an ordinary site visitor without access to the Wagtail admin.

Many thanks to xutao and Rajesh Murthy for reporting this issue. For further details, please see security advisory GHSA-jm43-45cc-q9rr.

Security fix: Formula injection via XLSX export

A vulnerability existed in Wagtail’s form builder export feature and other areas where untrusted user input may be exported to XLSX format, such as reports. An attacker could insert a string value which, when exported to XLSX format and opened in an external spreadsheet application, is interpreted as a formula. The impact of this is dependent on the spreadsheet application, but may include command execution or data filtration (most likely requiring further interaction from the user opening the spreadsheet).

Note that this vulnerability is distinct from the known issue of CSV formula injection, which is outside of Wagtail’s scope as it depends on the spreadsheet software providing its own interpretation of the file contents that has no basis in the CSV specification. In the XLSX case, the file exported by Wagtail explicitly identifies cells as containing formulae.

For further details, please see security advisory GHSA-g6cr-c9h3-6j2m.

Security fix: Improper restriction handling on Pages admin API for ancestor pages

The internal Pages admin API allowed authenticated users to access arbitrary contents of pages which are either a direct parent of a page the user has edit permission over, or a common ancestor of multiple pages the user has edit permission over. This could allow them to retrieve data that would otherwise be unavailable to them.

The vulnerability is not exploitable by an ordinary site visitor without access to the Wagtail admin.

Many thanks to Nguyen Dang Khoi for reporting this issue. For further details, please see security advisory GHSA-3xgh-xc8p-9h54.

Security fix: Stored XSS attack via v3 write API and third-party HTML editors

A cross-site scripting vulnerability existed in the handling of rich text input by the v3 API (released as a preview in Wagtail 8.0). An authenticated user with edit permission over pages or snippets could submit specially-crafted rich text content through the API that bypasses Wagtail’s sanitisation logic, allowing arbitrary HTML (including scripting) to be inserted into page or snippet content. This vulnerability is also exploitable via third-party rich text editors such as wagtail-hallo which use Wagtail’s legacy “Editor HTML” conversion logic. Sites that use the default Draftail rich text editor, and do not have the v3 API configured for write access, are not affected.

Many thanks to ma4ter0x22 and Alberto Fernandez-de-Retana for reporting this issue. For further details, please see security advisory GHSA-4pg7-87vv-vrh7.

Security fix: Improper restriction handling on API v3 pages endpoint

The v3 API (released as a preview in Wagtail 8.0) incorrectly allows authenticated users to access draft revisions of pages which are either a direct parent of a page the user has edit permission over, or common ancestors of multiple pages the user has edit permission over. This could allow them to retrieve data that would otherwise be unavailable to them.

The vulnerability is not exploitable by an ordinary site visitor without an API bearer token.

Many thanks to Longwei Wang, Cerebrovinny, Anass Bali, Muted Cybersecurity S.L. and Dexx-tech for reporting this vulnerability. For further details, please see security advisory GHSA-jrgm-fhgv-hmj9.

Bug fixes

  • Prevent duplicate meta schemas for models used in foreign keys (Sage Abdullah)

  • Generate best-effort schema for API v3 APIFields with a custom serializer that has type hints on to_representation (Sage Abdullah)

  • Prevent API v3 PATCH requests from discarding unpublished draft changes of live pages and snippets (Sage Abdullah)

  • Correctly handle tags passed as a list of names in API v3 write requests (Sage Abdullah)

  • Respect submitted order of child relation items in API v3 write requests (Sage Abdullah)

  • Respect custom primary keys in API v3 write schema for inline panels (Sage Abdullah)

  • Fix child relations with related_query_name or default accessor being ignored by API v3 (Sage Abdullah)

Maintenance

  • Suppress warnings generated under sphinx-llm 1.1.0 (Matt Westcott)

  • Fix erroneous workflow creation test failing on Django 5.2.18 / 6.0.9 / 6.1.2 (Sage Abdullah)