🔍 Read the full analysis: Patch, PHP2Shell, Root: A Two-Day Palo Alto Server Attack on ThorstenMeyerAI.com
Get hardware and tech essentials 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 says an automated attack rooted 24 Palo Alto Networks firewalls on October 23, two days after PHP2Shell fixes were released. The attacker then used two firewalls to enter a US government-affiliated network and compromise an internal management server. Dream found AI tools in the attacker’s environment but says it has no direct evidence they created the exploit.
Attackers rooted 24 Palo Alto Networks firewalls in an automated campaign on October 23, two days after PHP fixes for the flaws called PHP2Shell were released, according to Israeli security company Dream. Dream says the attackers then used two firewalls to enter a US government-affiliated network, compromise an internal firewall-management server and target Active Directory, the system used to manage access to accounts and resources.
Dream’s account, published October 26 and updated October 29, describes an attack against 240 selected firewalls. The systems were in the United States, Japan, Taiwan and South Korea. Dream says the attacker gained root access on 24 of them. Palo Alto Networks said its Unit 42 researchers independently verified the attacker’s tools and results.
The attack chained PHP2Shell vulnerabilities CVE-2026-109618 and CVE-2026-109619 with Fragnesia, a Linux kernel privilege-escalation flaw assigned CVE-2026-46300 and made public in May. Dream says PHP2Shell allowed commands to run as a restricted web-server user; Fragnesia then helped the attacker gain root access. The company says the chain, including installing a persistent back door, took less than 15 minutes.
Dream reports that 22 rooted firewalls were left with a hidden web shell. In the government-affiliated organization, the attacker compromised two firewalls before using a tunnel to reach an internal server that manages the organization’s firewalls. Dream places that server compromise at 11:35 UTC on October 24 and says the attacker subsequently targeted Active Directory.
Palo Alto Networks issued an advisory on October 21, but Dream says fixes for the affected PAN-OS 12.1 and 12.2 branches shipped between October 27 and 29. Until fixed software was available, Palo Alto’s stated mitigation was to keep the management interface off the public internet. The dates indicate that disclosure of a flaw and availability of a vendor fix were separate steps.
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
Exposed Management Pages Enabled Access
The reported compromise shows how an internet-accessible firewall management interface can turn a software vulnerability into a foothold inside an organization. Firewalls are meant to control network traffic, but their administrative interfaces can also provide powerful access if attackers reach them and gain elevated privileges.
The government-affiliated intrusion raises the stakes beyond the 24 edge devices. Dream says the attacker reached a server used to manage the organization’s firewalls and went after Active Directory, which can govern access across many systems. The account does not establish what data, if any, was taken, or whether the attacker gained control over additional internal systems.
The timeline also matters for organizations running affected versions: fixes for the relevant PAN-OS branches followed the advisory by several days. During that interval, limiting management access to trusted networks was a key defensive measure, according to the vendor information cited by Dream. The episode underlines the need to track both vulnerability disclosures and the release of usable patches.
Products Worth Considering
LG UltraGear 32GX870A-B 32-inch 4K OLED Gaming Monitor
Ledger Flex + Folio Crypto Hardware Wallet
Trezor Safe 7 Crypto Hardware Wallet
As an affiliate, we earn on qualifying purchases.
From May Kernel Flaw to October Attack
Fragnesia was disclosed in May with a public exploit and a kernel fix, according to the source account. The PHP project released fixes for PHP2Shell on October 21. Dream says the attacker had set up the offensive-security platform aiscan on October 8, with DeepSeek and Kimi models configured, before scanning for exposed PAN-OS management pages.
Dream says an initial search identified about 19,000 firewalls with management pages reachable from the internet. A custom scanner obtained version information from about 5,000, after which country and version filters narrowed the set to 240 targets. The 24 rooted devices are a result within that selected group; they should not be read as a count of all vulnerable or exposed firewalls.
Dream also found Claude Code installed on the attacker’s servers, alongside the configured models. Those observations help explain why the incident has been described as AI-assisted, but they do not establish that AI generated the exploit or ran the attack independently.
““no direct evidence” that AI tools wrote the exploit or scripts”
— Dream
AI Role and Intrusion Impact
The precise role of AI remains unconfirmed. Dream found Claude Code and configured DeepSeek and Kimi models on the attacker’s systems, but said it had no direct evidence that they wrote the exploit or scripts. The available account does not establish whether the tools were used for reconnaissance, coding, analysis or another task.
The full impact on the government-affiliated organization is also unclear. Dream reports that the attacker targeted Active Directory, but the source material does not say whether credentials were stolen, accounts were changed, data was accessed or extracted, or how long the attacker retained access. It does not identify the organization.
Although Palo Alto Networks said Unit 42 independently verified the tools and results, the source does not provide a detailed public breakdown of that verification. The number of affected systems beyond the 25 reported servers, and whether any other organizations were breached through this method, remain unknown.
Patch Affected PAN-OS Branches
Organizations using affected PAN-OS versions should apply the vendor’s relevant fixed builds and check whether management interfaces are exposed to the internet. Dream’s account cites fixes for the 12.1 and 12.2 branches released between October 27 and 29; administrators should confirm the applicable version and guidance with Palo Alto Networks rather than assume every system was covered by the same release.
Security teams should also review firewall and authentication logs for unexpected administrative activity, web shells, tunnels or access to internal management systems. Because the report describes attempts against Active Directory, checking for unusual account and directory changes is relevant to determining whether an intrusion extended beyond a firewall.
Dream and Palo Alto Networks may provide further technical findings, while affected organizations and public authorities may clarify the scope and impact. The source material does not report a final assessment of data loss or remediation inside the government-affiliated network, so those questions remain open.
Key Questions
What happened in the Palo Alto firewall attack?
Dream says an attacker used an automated campaign on October 23 to exploit exposed PAN-OS firewalls, gaining root access on 24 of 240 selected targets. The attacker later used two firewalls to reach an internal server at a US government-affiliated organization.
What are PHP2Shell and Fragnesia?
PHP2Shell refers to two PHP-related flaws, CVE-2026-109618 and CVE-2026-109619. Dream says the attacker combined them with Fragnesia (CVE-2026-46300), a Linux kernel privilege-escalation vulnerability disclosed in May, to move from limited command execution to root access.
Did AI create or operate the attack?
That has not been established. Dream found Claude Code installed and DeepSeek and Kimi configured on the attacker’s servers, but said it had no direct evidence that AI tools wrote the exploit or scripts and that the operation did not appear fully autonomous.
How many systems were compromised?
Dream reports 25 servers: 24 firewalls and one internal firewall-management server. The 24 firewall compromises came from a selected batch of 240 targets, not from every exposed or vulnerable PAN-OS device.
What should administrators do?
Apply the fixed PAN-OS version that matches the affected system and restrict management interfaces to trusted networks rather than the public internet. Review firewall, tunnel and account activity for signs of unauthorized access. This is security reporting, not financial advice; software and incident risks can change as new findings emerge.
Source: ThorstenMeyerAI.com
