Security
Headlines
HeadlinesLatestCVEs

Headline

GHSA-vc7j-h8xg-fv5x: matrix-appservice-bridge doesn't verify the sub parameter of an openId token exhange, allowing unauthorized access to provisioning APIs

Impact

A malicious Matrix server can use a foreign user’s MXID in an OpenID exchange, allowing a bad actor to impersonate users when using the provisioning API.

Details

The library does not check that the servername part of the sub parameter (containing the user’s claimed MXID) is the same as the servername we are talking to. A malicious actor could spin up a server on any given domain, respond with a sub parameter according to the user they want to act as and use the resulting token to perform provisioning requests.

Workarounds

Disable the provisioning API. If the bridge does not use the provisioning API, you are not vulnerable.

ghsa
#auth

Impact

A malicious Matrix server can use a foreign user’s MXID in an OpenID exchange, allowing a bad actor to impersonate users when using the provisioning API.

Details

The library does not check that the servername part of the sub parameter (containing the user’s claimed MXID) is the same as the servername we are talking to. A malicious actor could spin up a server on any given domain, respond with a sub parameter according to the user they want to act as and use the resulting token to perform provisioning requests.

Workarounds

Disable the provisioning API. If the bridge does not use the provisioning API, you are not vulnerable.

References

  • GHSA-vc7j-h8xg-fv5x
  • matrix-org/matrix-appservice-bridge@4c6723a

Related news

CVE-2023-38691: Merge pull request from GHSA-vc7j-h8xg-fv5x · matrix-org/matrix-appservice-bridge@4c6723a

matrix-appservice-bridge provides an API for setting up bridges. Starting in version 4.0.0 and prior to versions 8.1.2 and 9.0.1, a malicious Matrix server can use a foreign user's MXID in an OpenID exchange, allowing a bad actor to impersonate users when using the provisioning API. The library does not check that the servername part of the `sub` parameter (containing the user's *claimed* MXID) is the the same as the servername we are talking to. A malicious actor could spin up a server on any given domain, respond with a `sub` parameter according to the user they want to act as and use the resulting token to perform provisioning requests. Versions 8.1.2 and 9.0.1 contain a patch. As a workaround, disable the provisioning API.