Author here, we found it while building & documenting automated dependency assessment workflows for ZeroPath recently. Better-Auth made for a good test case given its popularity (300K weekly downloads).
The vulnerability is a logic error in how the API keys plugin determines user context when a userId is specified. Fix is in version 1.3.26. This is the kind of business logic flaw that traditional dependency vetting (stars, existing CVEs, reputation) doesn't catch. We're working on tooling to make these audits more practical at scale.
Hi, I'm Etienne, one of the cofounders @ ZeroPath.
We do not use traditional static analyzers; our engine was built from the ground up to use LLMs as a primitive. The issues ZeroPath identified in Joshua's post were indeed surfaced and triaged by AI.
The vulnerability is a logic error in how the API keys plugin determines user context when a userId is specified. Fix is in version 1.3.26. This is the kind of business logic flaw that traditional dependency vetting (stars, existing CVEs, reputation) doesn't catch. We're working on tooling to make these audits more practical at scale.