Are your secrets safe - How mobile applications are leaking millions of credentials
Where
PasswordsCon
BSides Las Vegas
Duration
51 min
Shows how hardcoded API keys and credentials end up inside Android and iOS apps, and how to find them.
Summary
The talk starts from the problem that machine-to-machine credentials such as API keys, certificates and service tokens are often hardcoded or left in the wrong place, because people handle them even though systems use them. Mobile apps make this worse, since a simple app usually connects to many third-party services, and each connection needs its own secret.
The speaker looks at public GitHub data, where a yearly GitGuardian report counted about 10 million validated secrets in 2023, and notes that mobile-specific files such as Android manifests, strings.xml and iOS plist files appear often in leaks. The talk then walks through downloading Android and iOS apps, decompiling the APK and unzipping the IPA, and scanning the extracted code. The live scan found valid Google API keys, Slack webhooks and other service credentials in a store app. A penetration test on a bank app is described, where a hardcoded S3 bucket credential exposed thousands of stored check images. Third-party studies of Play Store apps are also summarised, with a large share containing hardcoded credentials.
The takeaways are that secrets belong on the server side behind a login, since obfuscation, Base64 encoding and splitting keys do not protect them. Where a key must live in the app, it should be tightly scoped, restricted by IP or service, and paired with app signing. Teams should assume source code will be read, scan repositories and builds for secrets, and rotate credentials on a regular schedule.
Related talks
Talks·1 hr 1 min
Abraham Aranguren
Walks through a security and privacy audit of Hong Kong's LeaveHomeSafe contact tracing app and its serious flaws.
Main Stage·43 min
Ian Roos · Mark Trumpbour
Presents the 2026 Pwnie Awards ceremony, handing out mock prizes for the year's notable exploits, research, and vendor responses.
Track 3·45 min
Yelena Williams · Trevor Costanzi
Shows how blind SSRF is found, exploited through out-of-band callbacks, and defended with secure infrastructure defaults.


