A Milan-based startup has uncovered a full-takeover exploit in macOS, using ChatGPT as a research tool. But the company can't report the vulnerability to Apple — because the company's new submission cap blocks the filing.
How the exploit was found
The startup's security team was testing macOS internals with the help of ChatGPT, probing for weaknesses. They found a chain of bugs that, if exploited, would give an attacker complete control over a target Mac — no user interaction required beyond the initial compromise. The team confirmed the exploit works on the latest version of macOS.
ChatGPT helped the researchers accelerate their analysis, generating code snippets and suggesting attack vectors. The startup declined to name the specific macOS component involved, citing responsible disclosure concerns.
Apple's submission cap blocks the report
Apple recently introduced a cap on the number of vulnerability reports a single researcher or organization can submit through its official security portal. The startup had already hit that cap with earlier, less critical findings. When they tried to file the full-takeover exploit, the system rejected the submission.
The company says it has no way to escalate the issue outside the capped portal. Apple's security team does not accept reports via email or other channels for this type of bug. The startup is now stuck with a critical vulnerability it cannot responsibly disclose.
What the cap means for security research
Apple's submission cap is designed to prevent spam and low-quality reports from overwhelming its security team. But the policy also catches legitimate, high-severity findings. The Milan startup is not the first to hit the limit — other researchers have complained publicly about the cap blocking serious bugs.
Apple has not commented on this specific case. The company's security documentation states that researchers who exceed the cap must wait until the next quarter to submit new reports. That means the exploit could remain unreported for weeks or months.
What happens next
The startup is weighing its options. It could wait for the cap to reset, hoping no malicious actor discovers the same flaw in the meantime. Or it could publish the exploit details publicly — a move that would force Apple's hand but also put users at risk.
For now, the exploit remains a secret known only to the startup and, potentially, anyone monitoring ChatGPT's output. The company has not decided on a course of action. Apple's submission cap leaves them with no good choices.
