Patch the Planet: global response coordination as agents uncover vulnerabilities

Addressing critical flaws in PHP

Over the past two months, we led a coalition of maintainers, researchers, and security providers to address critical vulnerabilities in PHP, an open-source programming language used to run websites and web applications. Calif identified a critical flaw using GPT‑5.6 Sol as part of our Patch the Planet initiative, with Wiz also independently discovering and reporting both vulnerabilities. PHP’s maintainers led the fixes and release, with our teams and partners like Wiz, Cloudflare, and others contributing research, technical support, and coordination to prepare protections ahead of disclosure.

For organizations using PHP, the priority now is applying the fixes. We’re sharing a PHP version checker and we will be launching an accompanying Codex skill to help operators identify potentially vulnerable PHP versions and, alongside PHP’s and their providers’ guidance, determine which updates apply.

Given the severity of these flaws and PHP’s widespread use, we consider this one of the most impactful security issues in recent memory. It highlights how AI can uncover security flaws with internet-scale impact, but the same capabilities can also help attackers rapidly develop offensive tools and exploits.

To our knowledge, frontier AI found these vulnerabilities before attackers did, opening a Defender’s Window for maintainers and security providers to prepare fixes and protections ahead of disclosure.

To help more defenders use these capabilities, we’ve committed $1 billion in subsidized Daybreak access through Daybreak for Frontline Defenders, focused on organizations with limited security resources protecting essential services.

A proposed framework for the AI era

As significant as they are, these bugs are among several significant vulnerabilities we’ve already found and are working to fix, and we expect similar findings to increase sharply in the coming months. Drawing on our experience thus far, we’re sharing a proposed approach, appropriate for the AI era, to responding to critical vulnerabilities in widely used software, and we’re asking the security community to help improve this framework. It brings together our models, security expertise, and maintainers to help turn findings into protection and build a response that can scale as AI capabilities advance.

Getting a fix into use can mean adapting it for older releases, preparing provider mitigations, and updating systems across many organizations. A repeatable process helps those teams act quickly as more vulnerabilities are discovered.

Through Patch the Planet, we contributed security expertise and model access to support the PHP response, and worked with maintainers and research partners to test fixes and prepare guidance. We also helped coordinate with infrastructure providers on protections for affected systems and engaged government cyber agencies to help guidance reach critical organizations. This supported the maintainer-led response and helped other defenders prepare.

For some findings, researchers can work directly with the software’s maintainers to investigate the issue and develop a fix. Others need a wider response because they affect many products or because distributors and service providers need time to prepare updates and protections. In those cases, the proposed framework can help maintainers identify who needs to be involved and coordinate the work. The response can involve five stages: validate and fix the upstream vulnerability, give affected organizations time to act, coordinate with government partners to help protect critical systems, prepare updates across affected products, and publish guidance and support deployment once a patch is available.

Building more secure software

Powerful AI capabilities are advancing across the industry, and we expect AI to play a growing role in helping defenders find vulnerabilities, monitor systems, and respond to threats. We’re investing in models, tools, and partnerships to help defenders put those capabilities to work.

Our investments include subscriptions for all maintainers through Codex for OSS and direct sponsorship of projects such as the Rust Foundation for maintainers and technical work. Each project sets its own priorities. Reusable tests and security workflows can reduce the effort needed to handle future findings. Open-source projects can apply to Patch the Planet for hands-on support.

We’re also calibrating additional blocking classifiers for Daybreak to help reduce the potential for misuse. This involves difficult tradeoffs: safeguards intended to prevent abuse can also block legitimate security work. Broad access to capable models and effective protections against misuse are both critical. Our aim is to preserve the capabilities defenders need to find and fix vulnerabilities while limiting harmful use. We’ll continue refining that balance as capabilities advance and we learn from how these tools are used.

As models become more capable, the opportunity extends beyond finding and fixing individual vulnerabilities. We want to help engineers build more secure software from the start. Improvements to a widely used component can prevent bugs or limit their impact across many products and services.