Impact

The Pomerium user info page (at /.pomerium) unintentionally included serialized OAuth2 access and ID tokens from the logged-in user's session. These tokens are not intended to be exposed to end users.

This issue may be more severe in the presence of an XSS vulnerability in an upstream application proxied through Pomerium. If an attacker could insert a malicious script onto a web page proxied through Pomerium, that script could access these tokens by making a request to the /.pomerium endpoint.

Upstream applications that authenticate only the ID token may be vulnerable to user impersonation using a token obtained in this manner.

Note that an OAuth2 access token or ID token by itself is not sufficient to hijack a user's Pomerium session. Upstream applications should not be vulnerable to user impersonation via these tokens provided: - the application verifies the Pomerium JWT for each request, - the connection between Pomerium and the application is secured by mTLS, - or the connection between Pomerium and the application is otherwise secured at the network layer.

Patches Patched in Pomerium v0.26.1.

Workarounds None

For more information If you have any questions or comments about this advisory:

- Open an issue in pomerium/pomerium - Email us at security@pomerium.com

Credit to Vadim Sheydaev, aka Enr1g for reporting this issue.

Impact In distributed service mode, Pomerium's Authenticate service exposes pprof debug and prometheus metrics handlers to untrusted traffic. This can leak potentially sensitive environmental information or lead to limited denial of service conditions.

Patches v0.17.1

Workarounds Block access to /debug and /metrics paths on the authenticate service. This can be done with any L7 proxy, including Pomerium's own proxy service.

References https://github.com/pomerium/pomerium/pull/3212

For more information If you have any questions or comments about this advisory: Open an issue in Pomerium Email us at security@pomerium.com

Pomerium is an open source identity-aware access proxy. In affected versions changes to the OIDC claims of a user after initial login are not reflected in policy evaluation when using allowedidpclaims as part of policy. If using allowedidpclaims and a user's claims are changed, Pomerium can make incorrect authorization decisions. This issue has been resolved in v0.15.6. For users unable to upgrade clear data on databroker service by clearing redis or restarting the in-memory databroker to force claims to be updated.

Impact Some API endpoints under /.pomerium/ do not verify parameters with pomeriumsignature. This could allow modifying parameters intended to be trusted to Pomerium.

The issue mainly affects routes responsible for sign in/out, but does not introduce an authentication bypass.

Specific Go Packages Affected github.com/pomerium/pomerium/authenticate

Patches Patched in v0.13.4

For more information If you have any questions or comments about this advisory Open an issue in pomerium Email us at security@pomerium.com

Impact Using programmatic access on protected sites, one can get a signed login URL with pomeriumredirecturi set to an arbitrary URL. Then, if the user has already logged into Pomerium, they will be redirected to the specified pomeriumredirecturi with a JWT attached. This allows an outside attacker to get a signed login URL that, upon visiting it, will redirect a victim to the attacker’s site. This creates an issue of Open Redirect and, more seriously, JWT leakage.

With a leaked JWT, the attacker will be able to unveil the victim’s identity (.e.g. email address) by supplying the JWT to the authenticate service or verify.pomerium.com. In addition, if an application integrating Pomerium only verifies the iss claim and others but not the aud claim, the attacker will be able to access it as the victim.

Specific Go Packages Affected github.com/pomerium/pomerium/proxy

Patches Patched in Pomerium v0.13.4

For more information If you have any questions or comments about this advisory Open an issue in pomerium Email us at security@pomerium.com

SQL injection vulnerability in ajax.php in Gregarius 0.5.4 and earlier allows remote attackers to execute arbitrary SQL commands via the rsargs array parameter in an expgetFeedContent action.

Cross-site scripting (XSS) vulnerability in search.php in Gallarific allows remote attackers to inject arbitrary web script or HTML via the query parameter. NOTE: the provenance of this information is unknown; the details are obtained solely from third party information.

Gallarific does not require authentication for (1) users.php and (2) index.php, which allows remote attackers to add and edit tasks via a direct request. NOTE: the provenance of this information is unknown; the details are obtained solely from third party information.