Cyber: Thousands Of Public Google Cloud API Keys Exposed With Gemini...

New research has found that Google Cloud API keys, typically designated as project identifiers for billing purposes, could be abused to authenticate to sensitive Gemini endpoints and access private data.

The findings come from Truffle Security, which discovered nearly 3,000 Google API keys (identified by the prefix "AIza") embedded in client-side code to provide Google-related services like embedded maps on websites.

"With a valid key, an attacker can access uploaded files, cached data, and charge LLM-usage to your account," security researcher Joe Leon said, adding the keys "now also authenticate to Gemini even though they were never intended for it."

The problem occurs when users enable the Gemini API on a Google Cloud project (i.e., Generative Language API), causing the existing API keys in that project, including those accessible via the website JavaScript code, to gain surreptitious access to Gemini endpoints without any warning or notice.

This effectively allows any attacker who scrapes websites to get hold of such API keys and use them for nefarious purposes and quota theft, including accessing sensitive files via the /files and /cachedContents endpoints, as well as making Gemini API calls, racking up huge bills for the victims.

In addition, Truffle Security found that creating a new API key in Google Cloud defaults to "Unrestricted," meaning it's applicable for every enabled API in the project, including Gemini.

"The result: thousands of API keys that were deployed as benign billing tokens are now live Gemini credentials sitting on the public internet," Leon said. In all, the company said it found 2,863 live keys accessible on the public internet, including a website associated with Google.

The disclosure comes as Quokka published a similar report, finding over 35,000 unique Google API keys embedded in its scan of 250,000 Android apps.

"Beyond potential cost abuse through automated LLM requests, organizations must also consider how AI-enabled endpoints might interact with prompts, generated content, or connected cloud services in ways that expand the blast radius of a compromised key," the mobile security company said.

"Even if no direct customer data is accessible, the combination of inference access, quota consumption, and possible integration with broader Google Cloud resources creates a risk profile that is materially different from the original billing-identifier model developers relied upon."

Source: The Hacker News