The world of web security was recently shaken by a significant incident involving the flawed disclosure of critical vulnerabilities in PHP, a widely-used open-source scripting language that powers a substantial portion of the internet. This debacle centered around two severe security bugs, collectively named PHP2shell, which were inadvertently exposed ahead of their official announcement by PHP developers. This early leak led to a complex series of events implicating software supply chain security company Chainguard in what PHP developers called a “leak and disrupted disclosure.”
The vulnerability discovery was made by researchers at Calif, OpenAI, and Wiz between August and September, with these entities employing AI models to detect the flaws. These vulnerabilities, designated as CVE-2026-109618 and CVE-2026-109619, are particularly alarming because they allow unauthenticated attackers to execute code on servers running PHP, posing a severe threat to a vast number of websites and online applications.
The saga began unfolding when rumors about these vulnerabilities surfaced on online forums and social media, accompanied by inaccurate details about the release time and the severity of the issues. These circumstances forced PHP maintainers to advance the planned disclosure from Thursday to Wednesday, in an attempt to mitigate the fallout. Interestingly, prior to this, Calif inadvertently released a video that demonstrated a proof-of-concept exploit and revealed additional information about these vulnerabilities.
Chainguard, a company specializing in securing software supply chains, found itself in the midst of this controversy. On the day earmarked for the coordinated release, more than six hours before the scheduled time, Chainguard published built images containing the vulnerability fix. Chainguard's CEO, Dan Lorenc, asserted that the company's intention was not to leak the vulnerabilities. Instead, they were reacting to indications of active exploitation in the wild. Lorenc emphasized that their decision to release the fixes early was driven by a policy of coordinated vulnerability disclosure, which bends when there is evidence of active exploits.
The PHP developers, on their part, expressed deep frustration, stating that Chainguard's action disregarded the extensive efforts towards a coordinated disclosure, potentially endangering many users by making the fix available in binary form before an official release was made available. This gave attackers an undue advantage by studying the fix while other PHP users remained vulnerable.
Chainguard's CEO defended their actions in discussions with the media. He explained that after being tipped off about upcoming vulnerabilities in PHP, their team managed to independently identify the same security issues using AI tools. He recounted how the events escalated quickly – from rumors about the vulnerability to witnessing actual exploits, which prompted the preemptive release of fixes.
The technical aspects of the vulnerabilities involve issues with how PHP handles memory management and error handling, leading to remote code execution (RCE) risks. CVE-2026-109618, for instance, involves out-of-memory corruption and CVE-2026-109619 deals with stale state vulnerabilities. Both vulnerabilities have the potential to disrupt or manipulate services running on affected servers severely.
This incident has highlighted a critical tension in the realm of cybersecurity: the balance between timely disclosure to prevent active exploitation and the need to coordinate such disclosures carefully to maximize user preparedness and response effectiveness. While PHP developers felt Chainguard's actions were premature, Lorenc stood by his decision, reflecting on the challenge of handling potentially weaponized vulnerabilities in a highly connected digital ecosystem. Meanwhile, security researchers underscore the pressing need for immediate patching given the widespread use of PHP, which, according to research from Wiz, powers approximately 70 percent of all websites.
As the dust settles, this incident serves as a reminder of the complexities in orchestrating vulnerability disclosures and the multifaceted challenges cybersecurity professionals face in an era where exploits can be developed rapidly and deployed globally with ease. It also underscores the evolving landscape of cybersecurity, where the rapid advancement of AI and other technologies continually reshape the dynamics of offense and defense in digital spaces.
Read the full post on theregister.com


