GHSA-7j9p-67mm-5g87: LTI 1.3 Grade Pass Back Implementation has Missing Authorization Vulnerability

Problem

TL;DR: Any LTI tool that is integrated with on the Open edX platform can
post a grade back for any LTI XBlock so long as it knows or can guess the
block location for that XBlock.

In LTI 1.3, LTI tools can “pass back” scores that learners earn while using
LTI tools to the edX platform. The edX platform then stores those LTI
scores in a separate table. If the right conditions are met, these scores
are then persisted to the LMS grades tables.

LTI tools can create what are called “line items” on the edX platform. A
line item can be thought of as a column in a grade book; it stores results
for a specific activity (i.e. XBlock) for a specific set of users (i.e.
users in the course using the XBlock). A line item has an optional
resource_link_id field, which is basically the XBlock location. An LTI tool
can supply any value for this field.

An LTI tool submits scores to the edX platform for line items. The code
that uploads that score to the LMS grade tables determines which XBlock to
upload the grades for by reading the resource_link_id field of the
associated line item. Because the LTI tool could have submitted any value
for the resource_link_id field, this introduces the potential for a
nefarious LTI tool to submit scores for any LTI XBlock on the platform.

Impact

Any LTI tool that is integrated with on the Open edX platform can post a
grade back for any LTI XBlock so long as it knows the resource_link_id
(i.e. block location) for that XBlock.

The impact is a loss of integrity for LTI XBlock grades.

Patches

No available patch

Workarounds

No

References

• GHSA-7j9p-67mm-5g87
• https://nvd.nist.gov/vuln/detail/CVE-2023-23611
• openedx/xblock-lti-consumer@252f94b
• https://github.com/pypa/advisory-database/tree/main/vulns/lti-consumer-xblock/PYSEC-2023-21.yaml