# CVE-2026-55736: Ash Framework private argument injection in Ash.Changeset

Source: https://junglewise.ai/threats/cve-2026-55736-ash-framework-private-argument-injection-in-ash-changeset  
Publisher: Junglewise Threat Intelligence, https://junglewise.ai/threats  
Published: 2026-06-23. Updated: 2026-06-23  
License: CC BY 4.0 with a link back

CVE-2026-55736 · Severity: medium · CVSS 5.9 · Published 2026-06-23

Ash is a framework for building Elixir applications. A security flaw was found where 'private' arguments—data intended to be set only by the server—can be manipulated by end-users. If an application uses these private arguments for critical tasks like verifying a user's identity or determining who owns a record, an attacker could potentially bypass security rules or modify data they shouldn't have access to.

Technologies: [Erlang Ash](https://junglewise.ai/threats/technologies/ash). Vendors: [Erlang](https://junglewise.ai/threats/vendors/erlang).

## Quick answers

**Is CVE-2026-55736 exploited in the wild?** No. Junglewise Threat Intelligence has no report of CVE-2026-55736 being exploited in the wild.

**How severe is CVE-2026-55736?** Medium severity, with a CVSS score of 5.9, published 23 June 2026.

**Which products and versions are affected?** Erlang Ash >= 3.0.0, < 3.29.3.

**Is there a fix for CVE-2026-55736?** A patch was published on 2026-06-23: Fix committed to ash-project/ash repository Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.

**How is CVE-2026-55736 attacked?** It is attacked with local access, in a software library.

**Am I affected by CVE-2026-55736?** Check whether you run Erlang Ash 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

A vulnerability exists in the Ash framework's changeset building process where private action arguments (defined with 'public?: false') are not consistently stripped from untrusted input. In the standard changeset path (for_create, for_update, for_destroy), the filtering logic fails when parameter keys are provided as strings (binaries) rather than atoms, which is the default for user-supplied data. Additionally, the atomic and bulk update paths (fully_atomic_changeset/4) fail to strip private arguments entirely. An attacker can exploit this to inject values into arguments intended for internal server-side use, such as 'acting_user_id', leading to privilege escalation or integrity violations. The issue is fixed in version 3.29.3.

## Affected products

- ash-project ash >= 3.0.0, < 3.29.3

## Identifiers

- CVE-2026-55736
- GHSA-f4hc-ppw9-4hhw

## Timeline

- 2026-06-23: advisory: Advisory published by GitHub and Erlang Ecosystem Foundation
- 2026-06-23: patched: Fix committed to ash-project/ash repository

## References

- [https://repo.hex.pm/](https://repo.hex.pm/)
- [https://github.com/](https://github.com/)
- [https://cna.erlef.org/cves/CVE-2026-55736.html](https://cna.erlef.org/cves/CVE-2026-55736.html)
- [https://github.com/ash-project/ash/commit/d9b3100219b3ea86d73202bf7368c03a7688efea](https://github.com/ash-project/ash/commit/d9b3100219b3ea86d73202bf7368c03a7688efea)
- [https://github.com/ash-project/ash/security/advisories/GHSA-f4hc-ppw9-4hhw](https://github.com/ash-project/ash/security/advisories/GHSA-f4hc-ppw9-4hhw)

## Related threats

- [Ash Framework Denial of Service via Atom Exhaustion in Module Type](https://junglewise.ai/threats/cve-2026-34593-ash-framework-denial-of-service-via-atom-exhaustion-in-module)