PHP Ships Emergency Fixes for Two Critical RCE Flaws Found by AI Agents
The PHP development team released emergency security updates on 21 October 2026 to fix two critical vulnerabilities, CVE-2026-109618 and CVE-2026-109619, which could allow unauthenticated remote code execution (RCE, where an attacker runs commands on a server they do not control). Help Net Security reported the emergency releases, and PHP had scheduled its critical security release for 14:00 UTC that day.
Wiz Research disclosed the flaws. It published its write-up on 20 October under the title "PHP2Shell: AI uncovers critical RCE vulnerabilities." The write-up describes flaws in PHP's core engine that enable remote code execution and Heartbleed-style memory disclosure, in which an attacker can read fragments of a server's memory. According to Wiz, the first bug was found on 19 August 2026 using Wiz Atlas, Gemini 4 Argon and GPT-5.6 Sol. The remote code execution vulnerability was discovered entirely by AI agents, a detail Wiz's research team walked through in a follow-up discussion published the same day.
The exposure is large. PHP is the open-source language behind roughly 70% of websites, so a flaw in its core reaches a wide range of hosts, content management systems (CMS platforms such as WordPress-style sites) and custom applications. Many of these are maintained by small teams with limited patch cycles. Some organizations took their sites offline until the vulnerabilities could be patched, according to TechCrunch's coverage. Reuters reported on 22 October that hacking attempts surged as companies and governments rushed to patch, and that Cloudflare had seen millions of exploit attempts following the disclosure.
Wiz Research has monitored exploitation attempts since disclosure. In a follow-up analysis published on 27 October, titled "PHP2Shell: Exploitation Attempts and Remediation Trends," the company reported that none of the attempts have succeeded so far.
Google Cloud also published a response on 20 October. Wiz is owned by Google. Wiz's CTO of Exposure Management and AppSec, Tomer Schwartz, said the company worked alongside OpenAI, the research group Calif, and the PHP team to fix the bugs. Schwartz was interviewed by Rebecca Bellan for an episode of TechCrunch's Equity podcast on PHP2Shell and what the episode called the "coming vulnpocalypse." Wiz is also hosting a 20-minute technical threat briefing with its research team for organizations assessing their exposure.
The coordination is the part of this story that matters most to operators. A disclosure involving a core language runtime, multiple AI labs and a commercial security vendor has to move through an embargo (a period of confidentiality), patch development and downstream packaging before the details become public. The fixes then had to reach hosting providers and application owners quickly enough to matter. Wiz's account places discovery in August and public disclosure in late October, and the emergency patch window was short. The slowest step was the patch pipeline: vendor release, distribution to hosts, application-level testing, and restart.
In my view, the more durable question is what this does to the economics of vulnerability management. If AI agents can surface critical RCE bugs in a language this widely deployed, the bottleneck moves from finding flaws to deploying fixes across a long tail of systems that nobody has inventoried. Organizations that know exactly which PHP versions they run, and can roll out a patched build within hours, will likely absorb this with little drama. Those that cannot will face the same problem the next time the disclosure cycle turns over.
The broader context here is encouraging, with a caveat. The discovery route Wiz describes uses the same class of tooling the industry has been debating for two years, and it found bugs in shared infrastructure that the community could fix in coordination with the language maintainers. Monitoring has so far found no successful exploitation, and the patches are in the hands of the people who run the sites. The open question is whether the pace of AI-assisted discovery will outrun the slow, unglamorous work of patching. Operators, not the disclosure itself, will settle that.
