Summary
A recently disclosed vulnerability in PHP, identified as CVE-2026-1096-18 and CVE-2026-1096-19, poses a significant risk to a large portion of the internet, as PHP is a foundational technology for many web applications. This vulnerability, which has a critical CVSS score of 9.2, allows for arbitrary code execution due to a flaw in memory management. The issue, initially noted in 2020, became exploitable after researchers connected various components of the vulnerability, leading to a rush for organizations to patch their systems. However, the patch was not released until hours after proof-of-concept exploits began circulating online, raising concerns about the timing and effectiveness of the response.
The discussion highlights the implications of the vulnerability, comparing it to the infamous Log4j incident, where widespread exploitation occurred rapidly. Experts emphasize the urgency for organizations to assess their use of PHP and implement patches immediately, as the vulnerability could affect a vast array of websites and applications, from small personal blogs to major corporate platforms. The role of artificial intelligence in identifying and exploiting vulnerabilities is also noted, with AI tools enhancing the capabilities of both attackers and defenders in this fast-evolving landscape.
Defenders face a challenging situation, needing to balance the risks of downtime against the potential consequences of exploitation. Many organizations, including government entities, are opting to take their systems offline until they can confirm they are secure. The conversation underscores the necessity for continuous security testing and the importance of understanding the interconnectedness of web technologies, as many systems rely on PHP in ways that may not be immediately apparent. As the situation develops, organizations are urged to remain vigilant and proactive in their security measures.
Show Notes
In this episode, we break down one security story: PHP2Shell: two critical PHP RCE flaws reaching back to PHP 5.4 How a leaked demo video broke PHP's coordinated disclosure Why AI turns every public detail into an exploit shortcut What defenders should patch, check, and monitor right now We also cover Bishop Fox research on the true blast radius of an AI attack, and how far a manipulated agent can reach when it acts inside enterprise systems.
Security Headlines: PHP Issues Emergency Patches for Two Critical PHP2Shell RCE Flaws as Exploit Attempts Emerge, The Hacker News php2shell: A tale of two PHP remote vulnerabilities, Calif Frequently Asked Questions About PHP Remote Code Execution (PHP2Shell CVE-2026-109618 & CVE-2026-109619), watchTowr
Also mentioned: Workshop: iOS Game Hacking: From Zero to God Mode Virtual Session: Healthcare Security at Machine Speed: $17M Ransomware & AI Adversaries Blog: The True Blast Radius of an AI Attack
Join the conversation in the Bishop Fox Discord server and in the Bishop Fox subreddit.
More from Initial Access
In this episode, we break down four security stories: A cyberattack that hijacked Danfoss refrigeration controllers at hundreds of French supermarkets An OpenAI engineer charged over a recording device hidden in an executive conference room Meta Muse VM escape flaws and the dossiers Muse builds on users and non-users ARTEX AI and Claude Code in South Korean bank attacks, and Anthropic pulling its AI tests offline We also cover Bishop Fox research on a zero-click XSS in Siemens Teamcenter, how physical penetration testing protects confidential meetings, and how to scope an AI security assessment around real risk decisions. Security Headlines: Cyberattack shuts down refrigerators at French supermarkets, BleepingComputer FBI arrests Saudi OpenAI employee who allegedly planted recording device in conference room, CNBC Meta Rushed to Fix Muse ‘VM Escape' Vulnerability Soon Before Launch, 404 Media Meta’s Muse AI Agent Is Building a Dossier On You, TIME ARTEX AI, Claude agents used in cyberattacks on South Korean banks, BleepingComputer Anthropic takes risky AI tests offline after models breach govt sites, BleepingComputer
In this episode, we break down three security stories: CVE-2026-88779, a third exploited NetScaler zero-day labeled denial-of-service About 8 million Danish CPR records stolen through a company's lawful access Google pausing its OSS VRP after a flood of invalid AI-generated reports Plus, we sit down with Andrew Crotty of the Ginger Hacker Initiative to talk about how newcomers actually break into cybersecurity in 2026, from choosing a path to building a resume that gets past the first screen. We also cover Bishop Fox research on how the AI attack surface extends across your full stack and what offensive testing reveals about healthcare security. Security Headlines: Citrix patches NetScaler SAML zero-day exploited in attacks, BleepingComputer Hackers steal 8 million citizens’ records from Danish government database, TechCrunch Google halts open-source bug bounty program amid AI spam surge, BleepingComputer Also mentioned: Events: No Hat 2026, HackMex 2026 Blog: Why the AI Attack Surface Extends Your Stack Guide: Healthcare Security: A Guide to Offensive Testing Andrew Crotty’s The Ginger Hacker Initiative Discord server. Join the conversation in the Bishop Fox Discord server and in the Bishop Fox subreddit.
In this episode, we break down four security stories: A ShinyHunters PeopleSoft WAF bypass that hinges on one URL-encoded character The arrest of an alleged ShinyHunters leader days before the group's FBI claim Nearly 400,000 DC Medicaid records left readable behind a public report OpenAI scrapping GPT-6.1 Astra over scope, authorization, and honesty We also cover Bishop Fox research on three practical risk questions: unauthenticated root RCE in Check Point management, Kubernetes namespace takeover in Zilliz Attu, and why an 85 percent jump in CVE disclosures does not automatically mean an 85 percent jump in real-world risk. Security Headlines: Google Warns of ShinyHunters' Fresh Oracle PeopleSoft Campaign, SecurityWeek Dutch police arrest suspected member of group that claimed FBI hack, BBC DC Health Agency Exposes 400,000 Beneficiary Records, SecurityWeek OpenAI scraps rollout of new AI model over safety concerns, BBC Also mentioned: Workshop: Weaponizing CloudFormation (available in English and Spanish) Events: SecureWorld Dallas 2026, VetsinTech NatSec + Space Innovation Summit 2026 Blog: One Port to Root: Weaponizing Check Point Management CVE-2026-93616 Blog: Zilliz / Attu | 2.6.5 Blog: Separating Signal from Slop: Triaging CVEs in the Age of AI Security Research Join the conversation in the Bishop Fox Discord server and in the Bishop Fox subreddit.
This episode breaks down: A zero-click Exchange Server RCE triggered by a Visio attachment AI agents compromising online retailers for about $25 a target ShinyHunters' unverified claim that it breached the FBI OpenAI's six new misalignment disclosures Plus, we sit-down with Bishop Fox's Emilio Gallegos on MikroTick, a RouterOS exploit chain attackers used before defenders knew it existed. Security Headlines: Microsoft Patches Record 974 Vulnerabilities, Including Two Exploited Zero-Days, Security Week Autonomous AI Agents are breaking into hundreds of Online Retailers for $25 a target in an ongoing campaign, Gambit Security FBI investigating claim hackers have stolen details of all its agents, BBC OpenAI Discloses Six New Incidents of ‘Concerning’ A.I. Behavior, The New York Times Also mentioned: Workshop: Weaponizing CloudFormation (available in English and Spanish) Virtual Session: AI & Cybersecurity Priorities for 2027: Executive Chat Events: BSides Cleveland 2026, Ekoparty 2026, No Hat 2026, Wild West Hackin' Fest Deadwood 2026 Blog: MikroTrick: Inside the RouteOS Takeover Chain Blog: Unified Code, Unified Risks: Uncovering Vulnerabilities in .NET MAUI Applications Blog: Master Key Included: Detecting SolarWinds ARM CVE-2026-28326 Join the conversation in the Bishop Fox Discord server and in the Bishop Fox subreddit.
This episode covers an AI agent fleet that compressed two weeks of red team work into ten hours, a two-CVE chain giving attackers full admin on MikroTik routers, and a fraudulent government request that walked Revolut's verification checks straight out the door, plus a sit-down with Bishop Fox's Jon Williams on a critical NetScaler auth bypass and the detection tool built to prove you're patched.