🔍 Read the full analysis: Did AI Assist The PHP2Shell Attack On 25 Palo Alto Servers? on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Dream says an attacker used PHP2Shell and a Linux kernel flaw to gain root access on 24 Palo Alto Networks firewalls, then compromised an internal firewall-management server at a U.S. government-affiliated organization. Dream found AI tools in the attacker’s environment but says it has no direct evidence they wrote the exploit or scripts.
Attackers used PHP2Shell and a previously disclosed Linux kernel flaw to gain root access on 24 Palo Alto Networks firewalls, then compromise an internal management server at a U.S. government-affiliated organization, according to an investigation by Israeli security firm Dream. Dream found AI tools on infrastructure it attributed to the attacker, but said it has no direct evidence those tools created the exploit or attack scripts.
Dream’s account says the attacker began an automated campaign on the evening of October 23, targeting 240 PAN-OS firewalls. The company says 24 were rooted. On October 24, the attacker used access to two firewalls associated with the same U.S. government-affiliated organization to reach a server managing the organization’s firewalls. Dream reports that this 25th compromised server was followed by attempts against the organization’s Active Directory, which controls user accounts and access to network resources.
The reported attack chained two vulnerabilities. PHP2Shell, identified as CVE-2026-109618 and CVE-2026-109619, allowed commands to run on the firewall’s web server, according to Dream. The attacker then used Fragnesia (CVE-2026-46300), a Linux kernel privilege-escalation flaw disclosed in May, to obtain root privileges. Dream says the full chain, including installation of a persistent back door, took less than 15 minutes in some cases.
Palo Alto Networks said its Unit 42 researchers independently verified the attacker’s tools and results, according to the source account. Dream says 22 of the 24 rooted firewalls were left with hidden web shells. The reported targets were in the United States, Japan, Taiwan and South Korea. The source material does not identify the affected organizations by name.
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 Firewalls Were at Risk
The reported breach shows how a newly fixed software flaw can remain exploitable when vulnerable systems are reachable from the public internet and a second, already disclosed weakness can raise an attacker’s privileges. Dream’s account says the attacker moved from exposed firewall management pages to persistent access and then to an internal server tied to a government-affiliated network. That makes the incident relevant beyond the individual devices: firewall-management systems can provide a route toward other sensitive infrastructure.
The AI element is less firmly established than the intrusion itself. Finding AI tools on a server associated with an attacker does not, by itself, show that AI generated the exploit, selected targets or controlled the operation. Dream explicitly said it had no direct evidence of AI authorship and that the operation did not appear fully autonomous. The verified concern is the vulnerability chain and the reported compromises; the extent of AI’s contribution remains an assessment, not a confirmed technical finding.
For defenders, the immediate point is practical: management interfaces exposed to the internet can face risk while patches are being prepared or deployed. The source account says Palo Alto’s advisory appeared on October 21, while fixed builds for the affected PAN-OS 12.1 and 12.2 branches arrived between October 27 and 29. During that gap, restricting management access was a key protective measure.
Products Worth Considering
FLEXISPOT Standing Desk with Monitor Riser & Cable Management
Samsung 34-Inch ViewFinity S50GC Ultra-WQHD Monitor
Dell UltraSharp 27 Thunderbolt Hub Monitor (U2726DER)
As an affiliate, we earn on qualifying purchases.
Patch Timing and the Exploit Chain
PHP released fixes for the PHP2Shell flaws on October 21. Palo Alto Networks issued an advisory that day, but, according to the source material, fixed releases for the affected 12.1 and 12.2 PAN-OS branches were not available until October 27–29. Dream says the firewalls use a vendor-shipped PHP version, which meant organizations needed to follow Palo Alto’s product-specific guidance rather than assume that installing a general PHP update on a firewall would resolve the issue.
Dream also described a target-selection process. It says a search of internet-connected devices identified about 19,000 firewalls with management pages exposed online. A custom scanner obtained version information from about 5,000, after which country and version filters narrowed the campaign to 240 targets. These are figures from Dream’s investigation; the source does not provide an independent count or describe the full filtering method.
The second flaw in the chain, Fragnesia, had been publicly disclosed in May with an exploit available, according to Dream’s timeline and the cited vulnerability references. The attack therefore combined a recently disclosed product vulnerability with an older kernel weakness. The sequence matters: PHP2Shell reportedly provided an initial foothold, while Fragnesia enabled the move to administrator-level access.
““no direct evidence” that AI tools wrote the exploit or scripts”
— Dream
AI’s Role Remains Unverified
The available account does not establish whether AI generated any code, helped adapt known exploits, assisted with target selection or was simply present on the attacker’s systems. Dream found Claude Code installed and DeepSeek and Kimi configured in an offensive-security platform called aiscan, but that evidence alone does not show how those tools were used.
Other details also remain limited in the supplied reporting. The affected organizations have not been publicly named, and the extent of access to the government-affiliated network and Active Directory is not fully described. The material reports attacks against Active Directory but does not specify what data, accounts or systems, if any, were accessed or altered. The attacker’s identity and motive are also not established.
Dream’s figures and conclusions describe its investigation; Palo Alto Networks’ reported validation provides additional support for the tools and results, but the source material does not include a complete independent incident report or confirmation from the affected organization.
Patch, Restrict, and Investigate
Organizations using affected PAN-OS releases should follow Palo Alto Networks’ advisory and apply the fixed versions relevant to their systems. Until systems are patched, the source account says management interfaces should not be exposed to the public internet; access should be restricted to trusted networks and authorized administrators. Defenders should also review systems for the reported web shells and other persistence, rather than treating a patch alone as proof that a device was not compromised.
The U.S. Cybersecurity and Infrastructure Security Agency added both PHP2Shell vulnerabilities to its Known Exploited Vulnerabilities catalog on October 26 and, according to the source material, set a three-day deadline for federal agencies. Further public information from Palo Alto Networks, CISA or the affected organization could clarify the scope of the intrusion and the response. For now, the reported compromise is established in Dream’s account and is said to have been independently checked by Unit 42; the attacker’s use of AI remains unproven.
Key Questions
Were 25 Palo Alto firewalls compromised?
No. Dream reports root access on 24 PAN-OS firewalls and a separate internal firewall-management server, making 25 compromised servers in total. The 25th server was not described as another Palo Alto firewall.
Did AI write the PHP2Shell exploit?
That has not been established. Dream found AI tools on infrastructure it attributed to the attacker, but said it had no direct evidence that AI wrote the exploit or scripts.
Which vulnerabilities were used?
Dream says the attacker chained PHP2Shell, CVE-2026-109618 and CVE-2026-109619, with the Linux kernel privilege-escalation flaw Fragnesia, CVE-2026-46300.
What should PAN-OS administrators do?
Follow Palo Alto Networks’ security advisory and install the fixed release for the relevant branch. The source material also recommends keeping the management interface off the public internet and checking for signs of persistence or compromise.
Source: ThorstenMeyerAI.com
