SonicWall SMA 1000 attacks began two days after the patch, with the advisory unchanged
Attackers began exploiting SonicWall SMA 1000 flaw CVE-2026-102255 on October 8, two days after the patch. In June, Rapid7 found attackers on this product took credentials, session databases and MFA seeds.
Not a builder's beat, but builders have a standing stake in it.
Timeline: in June 2026, threat actor UTA0533 weaponized two other SMA 1000 flaws as zero-days. On October 6, SonicWall disclosed CVE-2026-102255 and urged customers to update. On October 8, the first exploitation attempts reached Defused's SMA decoys.
| Date | Event | Claim |
|---|---|---|
| June 2026 | UTA0533 weaponizes two other SMA 1000 flaws as zero-days | 10 |
| October 6, 2026 | SonicWall discloses CVE-2026-102255 and urges updates | 5 |
| October 8, 2026 | First exploitation attempts reach Defused's SMA decoys | 2 |
What happened
- According to Security Affairs, SonicWall's advisory says the vendor found no evidence of exploitation but recommends applying the fixes.
- Defused said the first exploitation attempts reached its SMA decoys on October 8, only two days after public disclosure.
- SonicWall has yet to update its advisory to reflect active exploitation of CVE-2026-102255.
- Previdian data shows 135 exploitation attempts against CVE-2026-102255 from two unique IP addresses located in the U.S.
- SonicWall disclosed CVE-2026-102255 on October 6 and urged customers to update 6210, 7210 and 8200v models to software versions 12.5.0-03082 and 12.4.3-03670.
Why it matters
SonicWall's advisory and the outside sensors tell different stories. According to Security Affairs, the advisory says the vendor found no evidence of exploitation and recommends applying the fixes [1]. Defused said the first exploitation attempts hit its SMA decoys on October 8, two days after SonicWall disclosed CVE-2026-102255 on October 6 [2][5]. Previdian has since recorded 135 attempts from two U.S. IP addresses [4].
Attempts, not confirmed compromises
The October figures are sensor data. Defused counted traffic against decoys, and Previdian counted attempts [2][4]. None of the reports describes a confirmed compromise of a customer appliance. Security Affairs lists one reported attacker address, 146.70.143[.]142 [14]. Horizon3.ai warned that successful exploitation could let attackers take over vulnerable devices [8].
The affected models are the 6210, 7210 and 8200v [5]. SonicWall says the flaw does not touch SSL-VPN on its firewalls or the SMA 100 Series [7]. There is no workaround, and the hotfix comes from the MySonicWall portal [6]. So the hotfix is the only control, and attempts against decoys began two days after it was released [6][2].
Two addresses is a narrow base: 135 attempts from two sources averages 67.5 each [36]. None of the sources attributes the October attempts to a named actor.
What June showed a foothold is worth
SMA 1000 was also exploited earlier this year. In June 2026, CVE-2026-15409 (CVSS 10.0) and CVE-2026-15410 (7.2) were used as zero-days by an actor tracked as UTA0533, which deployed Java web shells and the KNUCKLEBALL malware [10]. With October's flaw, that makes three SMA 1000 CVEs across two exploitation episodes this year [35]. The June attacks came from a named actor using zero-days, while October's so far are exploitation of a patched flaw from two addresses, and no report connects the two.
Rapid7 documented what the June foothold was used for. It said threat actors targeted internet-facing devices as a stealthy initial access vector, ran commands on the operating system and kept persistent access [11]. Then the theft:
"Once they established a foothold on the appliance, the actors systematically extracted high-value credentials, active session databases, and Time-Based One-Time Password (TOTP) multi-factor authentication (MFA) seed configurations," Rapid7 said [12]. It added: "With these harvested resources, the threat actors quickly shifted to lateral movement, pivoting from the compromised appliance directly into the internal corporate network." [13]
In our view, applying the patch does not undo theft that happened before it. A hotfix fixes the entry point. Credentials, session databases and TOTP seeds copied before it was applied are a separate problem, and the June account puts all three in the attacker's hands [12]. October's flaw is a different bug. SonicWall describes its impact as letting an attacker reach internal functionality and perform unauthorized operations [9]. No public report says whether anyone used it to get into a real appliance before October 6.
Bricksforge: the same question on a smaller footprint
Patchstack detected exploitation of Bricksforge's CVE-2026-85097 on October 7 at 9:47 p.m. UTC [15]. Versions through 3.1.8.9 are affected, 3.1.8.10 fixes the flaw, and no login is needed to exploit it [16]. Patchstack reported that roughly 6,000 websites use the plugin [27]. The reports do not include the date the patch shipped, so the gap between fix and attack cannot be computed. SecurityWeek notes that it is not uncommon for attackers to target WordPress flaws shortly after patches are released [34].
The bug is a validation gap between two steps. "Bricksforge validates a file's MIME type when it is first uploaded. When a form is later submitted, however, it trusts the file metadata the client sends," Patchstack said [17]. An attacker requests a nonce from an unauthenticated AJAX endpoint, uploads a GIF/PHP polyglot that passes the MIME check, then submits a form whose file path points at the image while the url field ends in a PHP extension. The plugin places the file's contents at the PHP location, and the code runs when that file is requested [18]. In Patchstack's representative request, the PHP path sits under /wp-content/uploads/bricksforge/tmp/ [19]. The exploitation activity uses alternate extensions such as .php5, .php7 and .phtml, case changes, and encoded extensions to get past filtering [20].
The sources disagree on the score. The Hacker News and SC Media report a CVSS of 10.0, and SecurityWeek reports 9.8 [26][25]. The gap changes nothing about who can attack, since the flaw needs no login [16].
Patchstack's telemetry shows 63 unique source IPs, with six accounting for roughly 46% of requests [21]. In one cluster, 44 IPs sent 56 requests with identical payloads within seconds of each other. In another, three IPs tried 25 variations of PHP extensions and encodings against the same temporary file [22]. Patchstack said: "Both patterns point to automated exploitation over rotating proxies or distributed infrastructure. They do not, on their own, attribute the activity to a single threat actor." [23]
Set that beside the SMA numbers. The first Bricksforge cluster averages about 1.3 requests per address [37], against 67.5 attempts per address on SMA [36]. These are different vendors' sensors over different windows, so they are not comparable rates. The shapes still differ: the SMA attempts are concentrated, and the Bricksforge cluster is spread thin.
The upgrade closes the upload path
Exploitation could let an attacker seize control of the site, deploy web shells, read sensitive data and establish persistent access [24]. Jacob Krell of Suzu Labs said: "From there, the attacker can change the site, access data available to WordPress, or leave behind access for later." [30] We think the logic matches the appliance case. The upgrade closes the upload path and does not, by itself, remove a file an attacker already wrote.
Jason Soroko of Sectigo gave the remediation step: "I'd verify the version and upgrade to 3.1.8.10 or 4.0.1 immediately," said Soroko [28]. He also said teams should search wp-content/uploads/bricksforge/tmp for unexpected PHP files and review Bricksforge form-submission logs [29]. That is the directory in Patchstack's sample request [19].
Edge devices and the record
Help Net Security's weekly roundup summarizes an interview with Rapid7's Christiaan Beek about the BPFDoor backdoor. According to that summary, its operators go after mail gateways and other edge devices that can't run endpoint agents [32]. BPFDoor is a different malware family. The constraint still applies to any appliance without an agent: there is no endpoint telemetry to read afterwards. In our view, that leaves the appliance's own records, and whatever an owner copied off it, as the evidence of device state.
The same roundup lists a NetScaler memory overflow bug, CVE-2026-88779, that CISA added to its Known Exploited Vulnerabilities catalog, and describes it as one that may cripple vulnerable ADCs and Gateways [33]. Its headline says FortiBleed "is still active" [31]. The roundup gives FortiBleed only that headline, with no CVE, affected versions or mechanism.
What the evidence supports
The case for treating remediation as proof of device state rests on two documented facts. One is Rapid7's June account of credential, session and MFA-seed theft on SMA 1000 [12]. The other is the persistent access Patchstack's reporting says Bricksforge exploitation can establish [24]. The reporting does not describe an attacker lockout mechanism on any of these products. And the October SMA numbers remain attempts logged by decoys and sensors [2][4].
As of The Hacker News and Security Affairs reporting, SonicWall had not updated its advisory to reflect active exploitation [3].
What to watch
- Whether SonicWall updates the CVE-2026-102255 advisory to confirm exploitation, and whether Previdian's count moves past 135 attempts or beyond two source addresses.
- Whether anyone attributes the October SMA 1000 attempts to UTA0533 or another named actor, or reports the credential, session and MFA-seed theft Rapid7 described in June.
- Whether Patchstack ties its Bricksforge clusters to a single actor, and whether owners report unexpected PHP files in the bricksforge/tmp directory.
Claim ledger
Ranked by verification strength, evidence, and original report placement.
- [1]
According to Security Affairs, SonicWall's advisory says the vendor found no evidence of exploitation but recommends applying the fixes.
- [2]
Defused said the first exploitation attempts reached its SMA decoys on October 8, only two days after public disclosure.
- [3]
SonicWall has yet to update its advisory to reflect active exploitation of CVE-2026-102255.
- [4]
Previdian data shows 135 exploitation attempts against CVE-2026-102255 from two unique IP addresses located in the U.S.
- [5]
SonicWall disclosed CVE-2026-102255 on October 6 and urged customers to update 6210, 7210 and 8200v models to software versions 12.5.0-03082 and 12.4.3-03670.
- [6]
The hotfix is available through the MySonicWall portal, and the vendor said there is no workaround.
- [7]
SonicWall said the flaw does not affect SSL-VPN running on SonicWall firewalls or the SMA 100 Series product line.
- [8]
Horizon3.ai warned that successful exploitation of the flaw could allow attackers to take over vulnerable devices.
- [9]
CVE-2026-102255 (CVSS 10) is a pre-authenticated SSRF in the SMA1000 Work Place interface; SonicWall says it could allow a remote attacker to direct the appliance to issue requests on their behalf and reach internal functionality and perform unauthorized operations.
- [10]
In June 2026, two other flaws in SMA 1000 (CVE-2026-15409, CVSS 10.0, and CVE-2026-15410, CVSS 7.2) were weaponized as zero-days by a threat actor dubbed UTA0533 to deploy Java web shells and a malware called KNUCKLEBALL.
- [11]
Rapid7 said it observed threat actors leveraging the June flaws to target internet-facing devices and use them as a stealthy initial access vector to execute commands on the operating system and ensure long-term, persistent access.
- [12]
"Once they established a foothold on the appliance, the actors systematically extracted high-value credentials, active session databases, and Time-Based One-Time Password (TOTP) multi-factor authentication (MFA) seed configurations," Rapid7 said.
- [13]
Rapid7 added: "With these harvested resources, the threat actors quickly shifted to lateral movement, pivoting from the compromised appliance directly into the internal corporate network."
- [14]
Security Affairs lists the reported attacker IP as 146.70.143[.]142.
- [15]
Patchstack said it detected exploitation efforts targeting the Bricksforge flaw on October 7, 2026, at 9:47 p.m. UTC.
- [16]
CVE-2026-85097 lets unauthenticated attackers upload and execute arbitrary PHP code on vulnerable WordPress sites; it affects Bricksforge versions up to and including 3.1.8.9 and is patched in 3.1.8.10.
- [17]
"Bricksforge validates a file's MIME type when it is first uploaded. When a form is later submitted, however, it trusts the file metadata the client sends," Patchstack said; in particular it does not validate the 'url' field inside the 'temporaryFileUploads' parameter.
- [18]
The exploit chain: request a nonce from the unauthenticated 'bricksforge_regenerate_nonce' AJAX endpoint; upload a GIF/PHP polyglot that passes the MIME check into the temporary upload directory; submit a form whose 'file' path points to the validated image but whose 'url' ends in a PHP extension; Bricksforge trusts the submitted metadata and places the file's contents at the PHP location, executing the code when requested.
- [19]
In Patchstack's representative request, the 'url' value is https://example.com/wp-content/uploads/bricksforge/tmp/img_example.php, in the same tmp directory as the validated GIF.
- [20]
To get around extension filtering, the exploitation activity uses alternate PHP extensions (.php5, .php7, .phtml), case manipulation (.PHP, .Php, .pHp) and encoded extensions.
- [21]
Exploitation activity has been recorded from 63 unique source IPs, with six of them accounting for roughly 46% of all requests.
- [22]
In one cluster, 44 IPs sent 56 requests with identical payloads within seconds of each other; in another, three IPs tested 25 variations of PHP extensions and encodings against the same temporary file.
- [23]
"Both patterns point to automated exploitation over rotating proxies or distributed infrastructure. They do not, on their own, attribute the activity to a single threat actor," Patchstack said.
- [24]
Successful exploitation could result in arbitrary PHP execution that an attacker can use to seize control of the WordPress site, deploy web shells, read sensitive data and establish persistent access.
- [25]
SecurityWeek lists CVE-2026-85097 with a CVSS score of 9.8.
- [26]
The Hacker News and SC Media list CVE-2026-85097 with a CVSS score of 10.0.
- [27]
Patchstack reported that Bricksforge is used by roughly 6,000 websites.
- [28]
"I'd verify the version and upgrade to 3.1.8.10 or 4.0.1 immediately," said Jason Soroko, senior fellow at Sectigo.
- [29]
Soroko said teams should also search wp-content/uploads/bricksforge/tmp for unexpected PHP files and review Bricksforge form-submission logs.
- [30]
"From there, the attacker can change the site, access data available to WordPress, or leave behind access for later," said Jacob Krell, senior director for secure AI solutions and cybersecurity at Suzu Labs.
- [31]
Help Net Security's weekly roundup is headlined "FortiBleed is still active"; the supplied item gives no CVE, versions or mechanism.
- [32]
In an interview, Rapid7's Christiaan Beek explains why operators of the BPFDoor backdoor go after mail gateways and other edge devices that can't run endpoint agents.
- [33]
CISA added NetScaler vulnerability CVE-2026-88779, a memory overflow bug that may cripple vulnerable NetScaler ADCs and Gateways, to its Known Exploited Vulnerabilities catalog.
- [34]
It is not uncommon for threat actors to target WordPress vulnerabilities shortly after patches are released, including plugin flaws and weaknesses affecting WordPress itself.
- [35]
Three SMA 1000 CVEs have come under exploitation across two episodes in 2026 (two in June, one in October).
- [36]
135 attempts from two IP addresses averages 67.5 attempts per address.
- [37]
Patchstack's first cluster averages about 1.3 requests per IP address.
Sources & coverage · 5 publishers
The reporting this story was synthesized from, earliest first. Every link goes to the original.
- helpnetsecurity.comHelp Net Security5d agoWeek in review: FortiBleed is still active, Patch Tuesday forecast
- securityweek.comIonut Arghire4d agoCritical SonicWall Vulnerability Exploited Days After Disclosure
- thehackernews.cominfo@thehackernews.com (The Hacker News)4d agoAttackers Exploit CVSS 10.0 Bricksforge WordPress Flaw Enabling Unauthenticated RCE
- thehackernews.cominfo@thehackernews.com (The Hacker News)4d agoCritical SonicWall SMA 1000 Flaw Targeted in 135 Exploitation Attempts After Patch
- securityweek.comEduard Kovacs4d agoBricksforge WordPress Plugin Vulnerability Exploited in the Wild
- securityaffairs.comPierluigi Paganini4d agoSonicWall Patches CVSS 10.0 SSRF Flaw as Attacks Begin
- scworld.comSteve Zurier4d agoCritical Bricksforge WordPress flaw exploited in attacks
