New analysis has pulled again the curtain on a “deficiency” in Google’s “Check in with Google” authentication move that exploits a quirk in area possession to realize entry to delicate knowledge.
“Google’s OAuth login would not defend in opposition to somebody buying a failed startup’s area and utilizing it to re-create e-mail accounts for former staff,” Truffle Safety co-founder and CEO Dylan Ayrey mentioned in a Monday report.
“And whilst you cannot entry previous e-mail knowledge, you should use these accounts to log into all of the totally different SaaS merchandise that the group used.”
The San Francisco-based firm mentioned the difficulty has the potential to place thousands and thousands of American customers’ knowledge in danger just by buying a defunct area related to a failed startup and gaining unauthorized entry to previous worker accounts associated to numerous functions like OpenAI ChatGPT, Slack, Notion, Zoom, and even HR programs.
“Essentially the most delicate accounts included HR programs, which contained tax paperwork, pay stubs, insurance coverage data, social safety numbers, and extra,” Ayrey mentioned. “Interview platforms additionally contained delicate details about candidate suggestions, presents, and rejections.”
OAuth, brief for open authorization, refers to an open normal for entry delegation, permitting customers to grant web sites or functions entry to their data on different web sites with out having to present their passwords. That is achieved by making use of an entry token to confirm the person’s id and permit the service to entry the useful resource the token is meant for.
When “Check in with Google” is used to check in to an utility resembling Slack, Google sends the service a set of claims concerning the person, together with their e-mail tackle and the hosted area, which might then be utilized to log customers into their accounts.
This additionally implies that if a service is solely counting on these items of data to authenticate customers, it additionally opens the door to a state of affairs the place area possession modifications might enable an attacker to regain entry to previous worker accounts.
Truffle additionally identified Google’s OAuth ID token features a distinctive person identifier – the sub declare – that would theoretically stop the issue, however that has been discovered to be unreliable. It is price noting that Microsoft’s Entra ID tokens embody the sub or oid claims to retailer an immutable worth per person.
Whereas Google initially responded to the vulnerability disclosure by stating that it’s meant habits, it has since re-opened the bug report as of December 19, 2024, awarding Ayrey a bounty of $1,337. It has additionally certified the difficulty as an “abuse-related methodology with excessive impression.”
Within the meantime, there are not any protections that downstream software program suppliers can take to guard in opposition to the vulnerability in Google’s OAuth implementation. The Hacker Information has reached out to Google for additional remark, and we’ll replace the story if we hear again.
“As a person, as soon as you have been off-boarded from a startup, you lose your potential to guard your knowledge in these accounts, and you’re topic to no matter destiny befalls the way forward for the startup and area,” Ayrey mentioned. “With out immutable identifiers for customers and workspaces, area possession modifications will proceed to compromise accounts.”