A credential sitting in plain sight inside a published app
Robert Wiggins found Wyndham's database password by decompiling its Android app and reading through the resulting code, a technique that works because a hardcoded credential is exposed to anyone who downloads and reverses the app, not only to someone with privileged internal access. Any mobile application a company publishes to a public app store is, from a security standpoint, publicly readable to anyone with the patience and tooling to decompile it, and that is exactly the exposure Wyndham's app had carried for an unknown period before Wiggins happened to look.
The credential he found granted direct access to an internal company database hosted on Salesforce, holding customer names, email addresses, rewards membership numbers, and JSON Web Tokens capable of keeping a customer's session signed in. Company files and internal documents were also reachable through the same access path, which means the exposure extended well past customer records alone into whatever internal material Wyndham had stored in that same Salesforce instance.
Why the exact scale is genuinely unknown, and that is the real problem
Wiggins reported that the database contained millions of entries but said he could not determine how many individual customer records were actually exposed within it. Wyndham's total rewards membership sits around 129 million people, and reporting on this incident has been careful to note that figure is the size of the membership program overall, not a count of how many of those members had their specific data exposed through this particular credential.
That gap between a database holding millions of entries and nobody being able to say precisely how many individuals were affected is itself a meaningful finding about Wyndham's own data governance practices. A company that cannot quickly answer how many records a given credential could access has a real visibility problem inside its own data architecture, one that exists independently of whether this specific exposure was ever actually exploited by an attacker, and that gap will outlast this particular incident unless it gets addressed directly.
What Wyndham did, and did not, commit to doing
Wyndham revoked the exposed credential only after TechCrunch's reporting team contacted its cybersecurity department directly, which means the company's own monitoring had not caught the exposure on its own before an outside journalist raised it. A spokesperson said the security team had taken steps to address the issue and found no evidence of unauthorized access, a reassuring claim that Wyndham has not backed with specifics about whether it has server access logs covering the full period the credential was live.
The company also declined to say whether it will notify affected customers directly, whether it plans to inform US regulators, or how it intends to permanently prevent a similar hardcoded-credential mistake in a future app release. It did provide an email address for reporting security vulnerabilities, but notably stopped short of committing to a formal, structured vulnerability disclosure program, which is precisely the kind of program that would have given Wiggins a clear, sanctioned channel to report this issue privately in the first place.
The disclosure gap that pushed a researcher to call a reporter instead
Wiggins reportedly turned to TechCrunch specifically because Wyndham offered him no clear way to report the flaw privately before going public with it, which is a recurring and avoidable failure mode across companies that have not invested in a formal bug bounty or coordinated disclosure program. A researcher who finds a credential like this one faces a real choice between staying quiet and hoping someone internal eventually finds it, or going public in a way that forces a fix but also announces the vulnerability's existence before it is confirmed resolved.
Companies that run structured disclosure programs, with a defined process, a named security contact, and often a financial reward, give researchers a reason to report privately rather than publicly, and they generally learn about their own vulnerabilities faster as a direct result. Wyndham's lack of such a program is not a minor process gap; it is the specific reason a straightforward hardcoded-credential finding became a public story rather than a quietly patched internal ticket.
Why AI-assisted decompilation is changing who finds these flaws
Wiggins used AI tools to help analyze the decompiled code in both this case and his earlier Applebee's finding, which points to a broader shift in who is capable of finding this class of vulnerability and how quickly. Decompiling and manually reading through an application's full code used to require specialized reverse-engineering expertise that limited how many people could realistically do this kind of audit at scale across the thousands of consumer apps published every year.
AI-assisted code analysis lowers that skill barrier meaningfully, which should worry any enterprise that has historically relied on obscurity, the assumption that reversing a mobile app is too much work for most attackers to bother with, as an informal layer of its own security posture. That assumption was already weak before AI tools made decompiled-code analysis faster and more accessible; it is now actively unsafe to rely on at all.
What this means for any enterprise shipping a mobile app with embedded credentials
Any organization that ships a consumer mobile app should treat a hardcoded database credential as a near-certain eventual discovery, not a low-probability risk, given how much easier AI-assisted reverse engineering has made this specific class of finding over the past year or two. The appropriate response is a mobile app security review that specifically checks for embedded secrets before every release ships, built into the standard release process itself, rather than a reactive scramble that only happens after a researcher or journalist surfaces one publicly.
Wyndham's specific failure, revoking a credential only after press contact rather than through its own internal detection, is the more instructive lesson here for other CIOs and CISOs than the underlying vulnerability itself. A security program's actual maturity shows up less in whether a mistake like this one ever happens at all, since mistakes happen everywhere even at well-resourced companies, and more in how quickly the organization's own systems catch it relative to how quickly an outside party eventually does instead.



