# CVE-2026-81643: ash-project ash_graphql authorization bypass in subscriptions

Source: https://junglewise.ai/threats/cve-2026-81643-ash-project-ash-graphql-authorization-bypass-in-subscriptions  
Publisher: Junglewise Threat Intelligence, https://junglewise.ai/threats  
Published: 2026-08-30. Updated: 2026-09-25  
License: CC BY 4.0 with a link back

CVE-2026-81643 · Severity: info · Published 2026-08-30

ash_graphql is a GraphQL library for the Elixir framework that allows real-time updates via subscriptions. This vulnerability allows an attacker who has access to GraphQL subscriptions to receive data updates for records they are not authorized to access, exposing sensitive information that should have been restricted.

Technologies: [Ash-Project Ash Graphql](https://junglewise.ai/threats/technologies/ash-graphql). Vendors: [Ash-Project](https://junglewise.ai/threats/vendors/ash-project).

## Quick answers

**Is CVE-2026-81643 exploited in the wild?** No. Junglewise Threat Intelligence has no report of CVE-2026-81643 being exploited in the wild.

**How severe is CVE-2026-81643?** Informational severity, published 30 August 2026.

**Which products and versions are affected?** Ash-Project Ash Graphql 1.4.0 to before 1.11.0.

**Is there a fix for CVE-2026-81643?** No fixed version is recorded yet. Check the vendor's advisory in the references below. If you run it, Junglewise drafts a mitigation that fits your setup and applies it once you approve, with senior security engineers behind it.

**How is CVE-2026-81643 attacked?** It is attacked over the network.

**Am I affected by CVE-2026-81643?** Check whether you run Ash-Project Ash Graphql at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

## Technical details

This is an authorization bypass vulnerability in AshGraphql.Subscription.Batcher's do_send/5 function. The first notification in a batch is correctly filtered with should_send?/1, which blocks unauthorized results (those with forbidden, not_found errors, or no error code). However, subsequent notifications in the same batch are fetched from the process dictionary and re-run through the pipeline without applying the authorization filter before being published. With batching enabled by default and a one-second batch window, two qualifying notifications are sufficient to leak unauthorized data. The fix applies the should_send?/1 filter to the entire batch, not just the first notification.

## Affected products

- ash-project ash_graphql 1.4.0 to before 1.11.0

## Timeline

- 2026-08-30: disclosed

## References

- [https://repo.hex.pm/](https://repo.hex.pm/)
- [https://github.com/](https://github.com/)

## Related threats

- [ash-project ash_graphql cross-session data exposure in subscriptions](https://junglewise.ai/threats/cve-2026-82367-ash-project-ash-graphql-cross-session-data-exposure-in)
- [ash-project ash_graphql query-complexity bypass](https://junglewise.ai/threats/cve-2026-81636-ash-project-ash-graphql-query-complexity-bypass)
- [ash-project ash_graphql incorrect authorization in subscriptions](https://junglewise.ai/threats/cve-2026-80223-ash-project-ash-graphql-incorrect-authorization-in-subscriptions)
- [ash-project ash_graphql information disclosure in error messages](https://junglewise.ai/threats/cve-2026-78693-ash-project-ash-graphql-information-disclosure-in-error-messages)