🔍 Read the full analysis: 25 Palo Alto Servers Hit By PHP2Shell: A Two-Day Path To Root on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Dream reported that an attacker exploited PHP2Shell and a Linux kernel flaw to gain root access on 24 Palo Alto Networks firewalls, then compromised a 25th server inside a U.S. government-affiliated network. Palo Alto Networks says its Unit 42 researchers independently verified the attack tools and results. Dream found AI tools on attacker systems but says it has no direct evidence they produced the exploit or scripts.
An attacker gained root access on 24 Palo Alto Networks firewalls and then used one to reach a 25th server inside a U.S. government-affiliated network, according to Israeli security company Dream. The reported intrusions began about two days after PHP released fixes for the vulnerabilities known as PHP2Shell; Palo Alto Networks says its Unit 42 researchers independently verified the attacker’s tools and results.
Dream’s October 26 report, updated October 29, says the attacker launched an automated campaign on the evening of October 23 against 240 selected firewalls. It reports that 24 were compromised with root privileges. The targets were in the United States, Japan, Taiwan and South Korea. Dream says 22 of the rooted firewalls were left with hidden web shells, while two belonged to the government-affiliated organization later penetrated.
The attacker chained PHP2Shell vulnerabilities CVE-2026-109618 and CVE-2026-109619 with Fragnesia, a Linux kernel privilege-escalation flaw tracked as CVE-2026-46300. Dream says the PHP flaws allowed code execution as a low-privilege account on the firewall; Fragnesia then enabled escalation to root. The company estimates the chain, including installation of a persistent back door, could take under 15 minutes. That timing and the attack details are Dream’s findings, not a public technical assessment reproduced here.
According to Dream’s timeline, the attacker compromised a second firewall belonging to the same organization on October 24, established a tunnel and persistence, then reached the organization’s internal firewall-management server at 11:35 UTC. Dream says the attacker subsequently targeted Active Directory, the system used to manage accounts and access. The report identifies the victim only as a U.S. government-affiliated organization; it does not name the entity.
Infographic 01 · PHP2Shell on Palo Alto firewalls
Two days from patch to root.
PHP fixed the bug on 21 October. Fixed firewall software for the attacked branches arrived almost a week later. The attack ran in between.
- 13–23 MayFragnesia goes public, kernel fix followsLinux privilege-escalation bug with a public exploit; CVE-2026-46300.
- 8 OctAttacker sets up an AI platformaiscan configured with DeepSeek and Kimi, per Dream’s timeline.
- 21 OctPHP fixes PHP2ShellPHP 8.5.12, 8.4.27, 8.3.36, 8.2.35. Palo Alto advisory the same day; PAN-OS fixes “upcoming”.
- 23 Oct, evening UTCBatch run: 240 targets, 24 rootedFound via device search engine, filtered to US, Japan, Taiwan, South Korea.
- 24 OctInto a government-affiliated networkTunnel, persistence, internal management server rooted (no. 25), Active Directory attacked.
- 26 OctDream publishes · CISA lists PHP2ShellFederal deadline 29 Oct.
- 27–29 OctFixed PAN-OS builds for 12.2 and 12.1CISA lists Fragnesia on 28 Oct; Unit 42 confirms Dream’s results on 29 Oct.
The fix for PHP existed. The fix for the firewall did not, yet. The open management page was the only door that mattered.Thorsten Meyer · my reading
Sources: Dream Research, 26 Oct 2026 (updated 29 Oct); php.net, 21 Oct 2026; Palo Alto Networks advisory (updated 29 Oct 2026); CISA KEV catalog, version 29 Oct 2026; oss-security, 13 May 2026; Unit 42, 29 Oct 2026. Day count: Thorsten Meyer.
Powered by Thorsten Meyer AI
Internet-Exposed Management at Risk
The reported campaign shows how a vulnerability in a firewall’s management software can become an entry point into the systems the device is meant to protect. In this case, Dream says the attacker moved from exposed management interfaces to root access and then used compromised infrastructure to reach an internal management server. That path could put an organization’s network controls and identity systems at risk, though the report does not establish what data, if any, was taken.
The timing also matters for administrators. PHP published fixes on October 21, but Dream says fixed PAN-OS builds for the affected 12.1 and 12.2 branches arrived between October 27 and 29. Palo Alto’s October 21 advisory described fixes as upcoming. During the intervening period, the vendor’s recommended protection was to keep the management interface off the public internet. The incident underlines that a software fix for an upstream component does not necessarily mean a vendor appliance can be patched immediately.
Dream found Claude Code installed on attacker systems and DeepSeek and Kimi configured in a platform called aiscan. Those observations do not prove the tools generated or executed the attack. The company said it had no direct evidence that AI wrote the exploit or scripts, and that the operation did not appear fully autonomous. The confirmed significance is the reported compromise; the degree to which AI contributed remains uncertain.
Products Worth Considering
Jackery Explorer 1000 v2 Portable Power Station
Bose SoundLink Flex (2nd Gen) Portable Bluetooth Speaker
Apple Watch SE 3 GPS 44mm Smartwatch
As an affiliate, we earn on qualifying purchases.
From PHP Fixes to PAN-OS Updates
PHP released updated versions on October 21, 2026, addressing the flaws Dream calls PHP2Shell. Palo Alto Networks issued an advisory that day, but according to the source account, the affected PAN-OS branches did not receive fixed builds until October 27-29. Dream says PAN-OS includes its own PHP version, making the vendor’s appliance updates relevant even after PHP’s project had shipped fixes.
The second vulnerability in the reported chain, Fragnesia (CVE-2026-46300), was disclosed in May 2026 with a public exploit and a kernel fix, according to the source material. Before the October 23 attacks, Dream says it identified about 19,000 PAN-OS firewalls with internet-facing management pages, obtained version information from about 5,000, and narrowed the pool to 240 targets based on country and software version. These figures describe Dream’s reported scanning and filtering; they are not a count of confirmed vulnerable or compromised devices.
The U.S. Cybersecurity and Infrastructure Security Agency added both PHP2Shell CVEs to its Known Exploited Vulnerabilities catalog on October 26 and set a three-day deadline for federal agencies, according to the source material. This federal directive does not by itself establish the status of systems outside the agencies covered by the requirement.
““no direct evidence” that AI tools wrote the exploit or scripts”
— Dream
Victim Impact and AI Role
Several details remain unresolved in the supplied reporting. The government-affiliated organization is not named, and the account does not specify whether the attacker accessed or removed data, how many people or systems were affected beyond the described servers, or whether the intrusion was contained. Dream says the attacker targeted Active Directory, but the material does not say what that activity achieved.
The evidence for AI assistance is also limited. Dream reports finding Claude Code and configured DeepSeek and Kimi models on attacker systems, but says it cannot directly link them to the exploit or scripts. It is not clear whether the tools were used for research, coding, or another purpose. Palo Alto Networks’ independent verification, as described in the source, concerns the tools and results; it does not resolve those questions about AI use or the victim’s ultimate losses.
The supplied account does not provide a complete patch-status count for exposed devices or confirm that all affected firewalls have been secured. The figure of 19,000 refers to internet-facing management pages found by Dream, not confirmed exposure to successful exploitation.
Patch Status and Incident Response
Administrators of affected PAN-OS versions should follow Palo Alto Networks’ current advisories and apply the fixed builds available for their branch. Where immediate patching is not possible, the source account identifies restricting management access so the interface is not reachable from the public internet as the interim protection. Organizations should also review firewall management activity and investigate signs of unauthorized persistence, using vendor guidance and their incident-response procedures.
Federal agencies face the deadline CISA set after adding the two PHP2Shell vulnerabilities to its exploited-vulnerabilities catalog. Further public updates may clarify whether the government-affiliated victim contained the intrusion, whether data was affected, and how broadly the campaign reached. The reporting provided does not state when those details will be released.
Key Questions
What was PHP2Shell?
PHP2Shell is the name used in the report for two vulnerabilities, CVE-2026-109618 and CVE-2026-109619, for which PHP released fixes on October 21, 2026. Dream says the attacker used them with the Fragnesia Linux kernel flaw to gain root access on targeted firewalls.
How many systems did the attacker reportedly compromise?
Dream reports root access on 24 Palo Alto firewalls, followed by access to one internal firewall-management server at a government-affiliated organization: 25 servers in total. The report does not name the organization.
Did AI create the exploit?
That is not established. Dream found AI coding and model tools on attacker systems but said it had no direct evidence they wrote the exploit or scripts. It also said the operation did not appear fully autonomous.
What should Palo Alto firewall administrators do?
Administrators should check Palo Alto Networks’ current advisory and install the fixed PAN-OS build for their branch. The source account says that, before fixes were available, the recommended protective measure was to keep the management interface off the public internet.
What is still unknown about the intrusion?
The victim’s identity, the extent of any data access or loss, the outcome of the Active Directory activity, and the role of the AI tools remain unclear in the supplied reporting. It also does not confirm how many exposed devices have since been patched.
Source: ThorstenMeyerAI.com
