CVE Database›Document archive›CVE-2026-49264

CVE-2026-49264

oauthlib: oauthlib: Unsafe JSONP callback injection in RevocationEndpoint allows arbitrary JavaScript response generation

Loading…

Severity

Moderate

Red Hat Product Security

6.1

CVSS 3.1

Review impact separately from score.

Published · September 28, 2026

Updated · November 1, 2026

Description

A flaw was found in oauthlib. When JSON with Padding (JSONP) support is enabled on the token revocation endpoint, the application fails to validate user-supplied callback parameters before reflecting them into the response. A remote attacker could exploit this flaw by submitting a crafted callback value, allowing the execution of arbitrary JavaScript code within the security context of client applications consuming the response.

Statement

This vulnerability is rated as Moderate impact because successful exploitation requires a non-default configuration alongside user interaction to execute injected script within the client application's origin. In typical deployments, JSONP processing is disabled by default within oauthlib, leaving standard configurations unaffected. Exposure is limited to custom services that explicitly activate JSONP handling on the token revocation endpoint and process untrusted callback inputs.

CVSS score applicability

The CVSS score(s) listed for this vulnerability do not reflect the associated product's status, and are included for informational purposes to better understand the severity of this vulnerability.

Mitigation

JSONP support on oauthlib's token revocation endpoint is opt-in and disabled by default. Deployments can mitigate this issue by not enabling it: ensure enable_jsonp=True is not passed when constructing oauthlib's RevocationEndpoint, so that user-supplied callback parameters are never reflected into JavaScript response bodies. Upstream removed the JSONP feature from RevocationEndpoint entirely in oauthlib 4.0.0.

CVSS / CWE

CVSS 3.1 · 6.1
CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:C/C:L/I:L/A:N

CWE-79 · Improper Neutralization of Input During Web Page Generation