CTIPilot
AI-generated · no human review · verify critical claims against the linked source. how it works →
‹Wed · 13 May 2026›
All daily briefs →
Daily brief · UTC day

Wednesday, 13 May 2026

12 verified findings from 1 run · 2 updates to prior coverage · the settled record for this UTC day, in the classic brief order.

Criticality
Kind
Topic
Region
TL;DR · the day in one read
  1. 01A Netlogon overflow gives unauthenticated attackers SYSTEM on domain controllers; Belgium's CCB reports exploitation that Microsoft has not confirmed. Microsoft's May 2026 updates fixed CVE-2026-41089, an unauthenticated stack overflow in Windows Netlogon that yields SYSTEM on domain controllers (CVSS 9.8), along with a DNS Client heap overflow (CVE-2026-41096), an SSO plugin privilege escalation (CVE-2026-41103) and a Dynamics 365 code injection (CVE-2026-42898). Belgium's CCB reported CVE-2026-41089 exploited on 2026-05-29; Microsoft said in June it had no evidence. →
  2. 02Exim 4.99.3 fixes an unauthenticated use-after-free in BDAT handling on GnuTLS builds (CVE-2026-45185). Exim 4.99.3, released on 2026-05-12, fixes CVE-2026-45185 (CVSS 9.8), a remotely reachable use-after-free in the GnuTLS backend: a client that sends a TLS close_notify before a BDAT body is complete and then one cleartext byte makes Exim write into its freed TLS transfer buffer, corrupting the heap. It affects Exim 4.97 to 4.99.2 built with GnuTLS that advertise STARTTLS and CHUNKING; GnuTLS is the default TLS library on many Debian-based distributions including Ubuntu, and the only mitigation is the upgrade. →
  3. 03Fortinet fixes unauthenticated code execution in the FortiAuthenticator API and the FortiSandbox web UI. Fortinet's advisories of 2026-05-12 fix CVE-2026-44277, improper access control on FortiAuthenticator API endpoints, and CVE-2026-26083, missing authorization in the FortiSandbox, FortiSandbox Cloud and FortiSandbox PaaS web UI; both let an unauthenticated attacker execute unauthorized code or commands through crafted requests (CVSSv3 9.8 base, 9.1 temporal). FortiAuthenticator is fixed in 6.5.7, 6.6.9 and 8.0.3 and FortiSandbox in 4.4.9 and 5.0.2, with Cloud and PaaS releases to upgrade to or migrate; Fortinet marked neither as known exploited at publication. →

01Active threats, incidents & disclosures2 items

NOTABLEupdatedNATOB2

Foxconn confirms a cyberattack on some North American factories after Nitrogen claims theft of 8 TB of customer project data

Foxconn said some of its factories in North America suffered a cyberattack, that its cybersecurity team activated its response mechanism to keep production and delivery going and that the affected factories were resuming normal production; a spokesperson declined to say how many factories were hit (The Record, 2026-05-12), and Foxconn declined to confirm that any customer's information was taken (The Register, 2026-05-12). The Nitrogen ransomware gang took credit on Monday 2026-05-11 (The Record, 2026-05-12), claiming 8 TB of data comprising more than 11 million files, with confidential instructions, internal project documentation and technical drawings related to projects at Intel, Apple, Google, Dell and Nvidia, among others (The Register, 2026-05-12). Workers at the Mount Pleasant, Wisconsin campus reported a network collapse from Friday 2026-05-01, according to The Cybersec Guru as quoted by 9to5Mac, and AppleInsider reports that a Houston, Texas facility also appears to have been affected (9to5Mac, 2026-05-12). Nitrogen's ransomware is believed to have been created with a builder based on the defunct Conti ransomware, and Barracuda describes the group as first observed in 2023 (The Record, 2026-05-12); Coveware found that a coding mistake in Nitrogen's ESXi encryptor overwrites four bytes of the public key, so files are encrypted with a corrupted key that nobody, the attackers included, can decrypt (Coveware, 2026-02-02).

The affected factories are currently resuming normal production

A spokesperson for the company confirmed the incident but declined to provide specifics on how many factories in North America were impacted.

The Record 2026-05-12

stolen 8 TB of data comprising more than 11 million files

Foxconn declined to confirm that these - or any - customers' information was hoovered up in the digital intrusion.

The Register 2026-05-12

Any ESXi encrypted files that are without viable backups, must be analyzed in conjunction with the corresponding malware that encrypted them to ascertain their status.

Coveware 2026-02-02
Correctionrun 2026-10-21T1148Z-auditaffected_productsbodyclassificationconfidenceevidenceheadlineprioritysourcessourcing_notesummarytechniquestitleentities

Foxconn's statement confirms that some of its North American factories suffered a cyberattack and are resuming normal production; it does not name sites, a start time or how long production stopped, and Foxconn declined to say how many factories were hit or to confirm that customer data was taken (The Record, 2026-05-12; The Register, 2026-05-12). Wisconsin, Ohio, Texas, Virginia, Indiana and several sites in Mexico are where Foxconn has factories, and Foxconn declined to say how many were impacted (The Record, 2026-05-12); the Mount Pleasant outage from 2026-05-01 comes from workers quoted by The Cybersec Guru, and Houston from AppleInsider (9to5Mac, 2026-05-12). Coveware's finding concerns Nitrogen's ESXi encryptor, whose bug leaves files unrecoverable even for the attackers, and its advice is to analyse encrypted files against the specific malware sample before assuming they can or cannot be recovered (Coveware, 2026-02-02).

incident13 May 05:00Zmulti-sourceOpen finding →
ROUTINEupdatedNATOB2

BWH Hotels (Best Western, WorldHotels, Sure Hotels): attackers had access to a guest-reservation web application for about six months

BWH Hotels, which operates more than 4,000 hotels including WorldHotels, Best Western Hotels & Resorts and Sure Hotels, told affected guests that it discovered the intrusion on 2026-04-22 and that the attackers had had access since 2025-10-14 to a web application holding guest names, email addresses, phone numbers and reservation details; payment and other financial information was not stored there, and how many guests are affected is unclear (SecurityWeek, 2026-05-12). The Register says the affected data goes back to 2025-10-14 and also includes home addresses, reservation numbers, dates of stay and special requests, that BWH took the application offline and revoked the unauthorised access, and that BWH has notified the appropriate regulatory agencies (The Register, 2026-05-11). No source states how the attackers got in, and no known cybercrime group has claimed the attack (SecurityWeek, 2026-05-12).

the intrusion was discovered on April 22, and an investigation showed that threat actors had access since October 14, 2025.

SecurityWeek 2026-05-12

identified the intrusion on April 22, but the affected data goes back to October 14, 2025.

The Register 2026-05-11

It’s unclear how many individuals were affected by the incident.

The hotel group seems concerned that the attackers may leverage the stolen data for scams and phishing.

SecurityWeek 2026-05-12
Correctionrun 2026-10-21T1148Z-auditbodyclassificationconfidenceentitiesevidenceheadlineprioritysourcessourcing_notesummarytitleverificationtechniques

BWH Hotels owns the WorldHotels, Best Western Hotels & Resorts and Sure Hotels brands, and the exposed fields are names, email addresses, telephone numbers, home addresses and reservation details, with no payment or bank details involved (The Register, 2026-05-11). SecurityWeek, quoting the guest emails, says the attackers had access from 2025-10-14 until the intrusion was discovered on 2026-04-22, while The Register says the affected data goes back to 2025-10-14; neither source discloses how the attackers got in or how many guests are affected (SecurityWeek, 2026-05-12), and BWH says only that it notified the appropriate regulatory agencies (The Register, 2026-05-11).

incident13 May 05:00Zsingle-source · victim disclosureOpen finding →
HIGHCVE-2026-41089 +3exploitedupdatedNATOA3

Microsoft May 2026 Patch Tuesday: Netlogon pre-auth RCE on domain controllers (CVE-2026-41089), later reported exploited by Belgium's CCB, leads four critical fixes

Microsoft's May 2026 Patch Tuesday fixed 118 vulnerabilities by Tenable's count, 16 rated critical, with none exploited or publicly disclosed at release (Tenable, 2026-05-12) (Krebs on Security, 2026-05-12); Help Net Security counts more than 120 (Help Net Security, 2026-05-12) and ZDI 138 (ZDI, 2026-05-12). The one that matters most is CVE-2026-41089 (CVSS 9.8), a stack-based buffer overflow in Windows Netlogon that lets an unauthenticated attacker execute code over the network (Microsoft MSRC); on a domain controller that means code execution as SYSTEM (CERT-EU, 2026-06-10). CVE-2026-41096 (CVSS 9.8) is a heap-based buffer overflow in the Windows DNS Client reachable through a malicious DNS response from an attacker able to influence replies (ZDI, 2026-05-12), which Microsoft assesses as less likely to be exploited (Krebs on Security, 2026-05-12). CVE-2026-41103 (CVSS 9.1) is an elevation of privilege in the Microsoft SSO Plugin for Jira and Confluence rated "Exploitation More Likely" (Tenable, 2026-05-12), and CVE-2026-42898 (CVSS 9.9) is a code-injection flaw in on-premises Dynamics 365 that requires an authenticated attacker (ZDI, 2026-05-12).

Contradiction: Belgium's CCB states that CVE-2026-41089 is exploited, on information from "trusted partners", while Microsoft told BleepingComputer it had no evidence to support the claim (CCB, 2026-05-29) (BleepingComputer, 2026-06-01).

CVE-2026-41089 (Windows Netlogon) is now actively exploited in the wild.

CCB 2026-05-29

Stack-based buffer overflow in Windows Netlogon allows an unauthorized attacker to execute code over a network.

Microsoft MSRC
Updaterun 2026-06-02-8af85d01actionscvesevidenceimmediate_actionpriorityregionssectorssourcestagsbody

Belgium's Centre for Cybersecurity updated its May Patch Tuesday advisory on 2026-05-29 to say that CVE-2026-41089 is now actively exploited in the wild (CCB, 2026-05-29), and BleepingComputer and Help Net Security reported it on 2026-06-01 (BleepingComputer, 2026-06-01) (Help Net Security, 2026-06-01). CCB attributes the information to trusted partners, and Microsoft said it had no evidence to support the claim while recommending the security updates (BleepingComputer, 2026-06-01). CERT-EU relayed CCB's statement in its advisory of 2026-06-10 (CERT-EU, 2026-06-10).

Correctionrun 2026-10-11T1307Z-audittitleheadlinesummaryimmediate_actionsectorstechniquesaffected_productscvessourcesevidenceclassificationactionsbodypriority

CCB's exploitation statement dates from 2026-05-29 and rests on information from trusted partners; Microsoft told BleepingComputer it had no evidence to support it (CCB, 2026-05-29) (BleepingComputer, 2026-06-01). Microsoft says 16 of the May CVEs, CVE-2026-41089 and CVE-2026-41096 among them, were found with its MDASH agentic system, and places the DNS Client flaw in dnsapi.dll (Microsoft Security Blog, 2026-05-12); no source says that flaw runs code as NetworkService. CVE-2026-41103 is an elevation of privilege, and CVE-2026-42898 requires an authenticated attacker (Tenable, 2026-05-12) (ZDI, 2026-05-12). The priority moves from critical to high: the exploitation claim rests on one national CERT's partner information, Microsoft said in a statement BleepingComputer added on 2026-06-02 that it had no evidence, and its advisory lists the flaw as not exploited (Microsoft MSRC); patching domain controllers remains the action.

vulnerability13 May 05:00Zmulti-sourceOpen finding →
NOTABLEupdatedNATOA2

CERTFR-2026-AVI-0572: Centreon April 2026 bulletin, SQL injection in eight components, command injection in MBI and XSS in MAP

Centreon published its April 2026 monthly security bulletin for Centreon Infra Monitoring on 2026-05-12: SQL injection through string concatenation across legacy PHP files in Anomaly Detection, Auto Discovery, AWIE, BAM, DSM, License Manager, MAP and Open Tickets (CVSS 8.8 each), command injection through a dos2unix exec call in MBI (CVSS 8.8) and an XSS in MAP through altered SVG images (CVSS 6.8), each listed with no reference identifier (Centreon, 2026-05-12). CERT-FR relayed it as CERTFR-2026-AVI-0572 with remote code execution, SQL injection and XSS as the risks, and lists no CVE identifiers (CERT-FR, 2026-05-12). The first fixed releases are Anomaly Detection 24.10.10 and 25.10.5, Auto Discovery 24.10.13 and 25.10.5, AWIE 24.10.6 and 25.10.6, BAM 24.10.14 and 25.10.5, DSM 24.10.6 and 25.10.3, License Manager 24.10.9 and 25.10.4, MAP 24.10.17 and 25.10.5, MBI 24.04.15, 24.10.17 and 25.10.5, and Open Tickets 24.10.11 and 25.10.6 (CERT-FR, 2026-05-12). Neither source states whether exploitation needs authentication or reports exploitation as of 2026-05-12 (Centreon, 2026-05-12; CERT-FR, 2026-05-12).

SQL injection via string concatenation across legacy PHP files

Command injection via dos2unix exec

XSS in images by altering the SVG

Centreon security bulletin 2026-05-12

MBI versions 24.04.x before 24.04.15 (translated from French)

CERT-FR CERTFR-2026-AVI-0572 2026-05-12
Correctionrun 2026-10-21T1148Z-auditaffected_productsbodyclassificationentitiesevidenceheadlinesourcessourcing_notesummarytechniquestitle

Centreon's bulletin lists every fix with no reference identifier and CERT-FR's advisory has no CVE section, so no CVE identifiers exist for this release (Centreon, 2026-05-12; CERT-FR, 2026-05-12). The flaws are SQL injection through string concatenation in legacy PHP files of eight components, command injection through a dos2unix exec call in MBI and an XSS through altered SVG images in MAP, each with its own fixed release per branch (CERT-FR, 2026-05-12). The Fortinet, Exim, Microsoft and SAP vulnerabilities patched the same week are separate advisories, not part of this Centreon release, and neither Centreon advisory describes Centreon's deployment in public-sector networks or the monitored-host credentials it holds (CERT-FR, 2026-05-12).

vulnerability13 May 05:00Zmulti-sourceOpen finding →
NOTABLEupdatedNATOA2

CERTFR-2026-AVI-0564: SPIP before 4.4.14, two remote code execution flaws in the private area and, on certain nginx configurations, the public area

SPIP's maintenance team released SPIP 4.4.14 on 2026-05-12 to fix two security flaws, a remote code execution (RCE) flaw in the private area and an RCE flaw in the public area that is limited to the use of certain nginx configurations, reported by Louka Jacques-Chevallier (Laluka), Baptiste Rolando, Arthur Deloffre (Vozec), jvoisin and Cezame (SPIP, 2026-05-12). The bulletin states that "these flaws are not covered by the security screen" (translated from French), so SPIP's built-in screen gives no interim protection; it gives no technical detail beyond the nginx condition, and the release's change list also includes better sanitisation of HTTP_HOST and a fix to REMOTE_USER handling under nginx (SPIP, 2026-05-12). CERT-FR's CERTFR-2026-AVI-0564 lists SPIP versions before 4.4.14 as affected with remote arbitrary code execution as the risk, refers to the SPIP bulletin for the fix and names no CVE identifiers (CERT-FR, 2026-05-12).

an RCE-type flaw in the public area, limited to the use of certain nginx configurations (translated from French)

These flaws are not covered by the security screen. (translated from French)

SPIP security bulletin 2026-05-12

SPIP versions before 4.4.14 (translated from French)

CERT-FR CERTFR-2026-AVI-0564 2026-05-12
Correctionrun 2026-10-21T1148Z-auditaffected_productsbodyclassificationentitiesevidenceheadlinesourcessourcing_notesummarytechniquestitle

SPIP's bulletin describes the public-area flaw only as limited to certain nginx configurations and names no exploit pattern, so there are no template or gadget signatures in access logs and no specific reverse-proxy location rule to audit for (SPIP, 2026-05-12). Its statement that "these flaws are not covered by the security screen" (translated from French) means the screen does not block them and the upgrade to 4.4.14 is the only fix; the bulletin counts two flaws (SPIP, 2026-05-12), and neither it nor CERT-FR gives figures on SPIP's use by ministries, universities or Swiss cantons (CERT-FR, 2026-05-12).

vulnerability13 May 05:00Zmulti-sourceOpen finding →
HIGHCVE-2026-45185updatedNATOA1

CVE-2026-45185 "Dead.Letter": Exim use-after-free in BDAT/CHUNKING on GnuTLS builds, fixed in 4.99.3

The Exim maintainers published EXIM-Security-2026-05-01.1 (CVE-2026-45185, Critical, CVSS 9.8) on 2026-05-12 after a report from Federico Kirschbaum of XBOW on 2026-05-01: when a client sends a TLS close_notify alert before a BDAT body transfer is complete and follows it with a final byte in cleartext on the same TCP connection, Exim writes into a buffer already freed during the TLS teardown and corrupts the heap, and an attacker only needs to establish a TLS connection and use the CHUNKING (BDAT) extension (Exim advisory, 2026-05-12). Exim 4.97 up to and including 4.99.2 built with USE_GNUTLS=yes are affected, OpenSSL builds are not, and 4.99.3 fixes the flaw by resetting the input-processing stack when a TLS close notification arrives during a BDAT transfer; there is no known mitigation other than upgrading (Exim advisory, 2026-05-12). XBOW's write-up places the freed memory in Exim's TLS transfer buffer, into which the still-active BDAT receive wrapper's call to ungetc() writes a single newline, and notes that GnuTLS is the default TLS library on many Debian-based distributions, including Ubuntu (XBOW, 2026-05-12). During the disclosure window an LLM guided step by step by XBOW Native produced working exploits against non-PIE builds, first without ASLR and then with ASLR, the latter running its payload through Exim's ${run} expansion, which forks bash; against the default PIE build no full exploit was reported: the researcher's side got as far as leaking a stack address, while the LLM never got as far as a leak (XBOW, 2026-05-12).

An attacker only needs to be able to establish a TLS connection and use the CHUNKING (BDAT) SMTP extension.

This vulnerability only impacts builds that use USE_GNUTLS=yes. Builds using OpenSSL or other TLS libraries are not affected.

There is no known mitigation other than upgrading.

Exim security advisory EXIM-Security-2026-05-01.1 2026-05-12

the default TLS library on many Debian-based distributions, including Ubuntu

XBOW Native had cracked the second challenge, ASLR on, still no-PIE.

The LLM kept working for a while longer and never even got as far as a leak.

XBOW 2026-05-12
Correctionrun 2026-10-21T1148Z-auditaffected_productsbodyclassificationcvesevidenceheadlinesourcessourcing_notesummarytechniquestitle

The LLM-built exploits covered non-PIE builds only, first without ASLR and then with ASLR; against the default PIE build the researcher's side got as far as leaking a stack address and the LLM never reached a leak, so no full exploit was reported (XBOW, 2026-05-12). The Exim maintainers state there is no known mitigation other than upgrading, so suppressing the CHUNKING advertisement is not a vendor-endorsed workaround, and their advisory, not XBOW's post, sets the CVSS 9.8 and the affected range of 4.97 up to and including 4.99.2 on builds using GnuTLS, fixed in 4.99.3 (Exim advisory, 2026-05-12). Exposure further requires STARTTLS and CHUNKING to be advertised (oss-security, 2026-05-12).

vulnerability13 May 05:00Zmulti-sourceOpen finding →
HIGHCVE-2026-44277 +1updatedNATOA2

CVE-2026-44277 and CVE-2026-26083: unauthenticated code execution in the Fortinet FortiAuthenticator API and the FortiSandbox web UI

CVE-2026-44277 is an improper access control flaw (CWE-284) on FortiAuthenticator's API endpoints that may let an unauthenticated attacker execute unauthorized code or commands through crafted requests; it affects FortiAuthenticator 6.5.0 to 6.5.6, 6.6.0 to 6.6.8, 8.0.0 and 8.0.2, is fixed in 6.5.7, 6.6.9 and 8.0.3, does not affect FortiAuthenticator Cloud, and can be worked around by disabling API access on exposed interfaces under Network, Interfaces, Access Rights (Fortinet FG-IR-26-128). CVE-2026-26083 is a missing authorization flaw (CWE-862) in the web UI of FortiSandbox, FortiSandbox Cloud and FortiSandbox PaaS with the same effect through HTTP requests; FortiSandbox is fixed in 4.4.9 and 5.0.2, FortiSandbox Cloud 5.0 in 5.0.6 and FortiSandbox PaaS 4.4 and 5.0 in 4.4.9 and 5.0.2, while FortiSandbox Cloud 23 and 24 and PaaS 21.3 to 23.4 must migrate to a fixed release (Fortinet FG-IR-26-136). Fortinet discovered both internally, rates each Critical with a CVSSv3 base score of 9.8 and a temporal score of 9.1, and marked neither as known exploited when it published on 2026-05-12 (Fortinet FG-IR-26-128; Fortinet FG-IR-26-136). ENISA's EUVD records carry the 9.1 figure and also list FortiAuthenticator 6.4.0 to 6.4.10 and FortiSandbox 4.2.1 to 4.2.8 as affected (EUVD-2026-29729; EUVD-2026-29550), branches for which Fortinet's advisories, as read on 2026-10-21, name no fixed release (Fortinet FG-IR-26-128; Fortinet FG-IR-26-136).

in FortiAuthenticator may allow an unauthenticated attacker to execute unauthorized code or commands via crafted requests.

FortiAuthenticator Cloud is not impacted by the issue and hence customers do not need to perform any action.

Fortinet PSIRT FG-IR-26-128 2026-05-12

in FortiSandbox, FortiSandbox Cloud and FortiSandbox PaaS WEB UI may allow an unauthenticated attacker to execute unauthorized code or commands via HTTP requests.

Fortinet PSIRT FG-IR-26-136 2026-05-12
Improvementrun 2026-10-21T1148Z-auditaffected_productsbodyclassificationcvesevidenceheadlinesourcessourcing_notesummarytechniquestitleevent_date

Fortinet rates both flaws at a CVSSv3 base score of 9.8 with a temporal score of 9.1 (Fortinet FG-IR-26-128; Fortinet FG-IR-26-136), and 9.1 is the figure ENISA's EUVD records carry (EUVD-2026-29729). The EUVD records also list FortiAuthenticator 6.4.0 to 6.4.10 and FortiSandbox 4.2.1 to 4.2.8 as affected (EUVD-2026-29550), branches for which Fortinet names no fixed release, and FortiSandbox Cloud 23 and 24 and PaaS 21.3 to 23.4 must migrate to a fixed release (Fortinet FG-IR-26-136). The mitigation Fortinet gives until FortiAuthenticator is upgraded is to disable API access on exposed interfaces under Network, Interfaces, Access Rights (Fortinet FG-IR-26-128).

vulnerability13 May 05:00Zmulti-sourceOpen finding →
NOTABLECVE-2026-34263 +1updatedNATOA2

CVE-2026-34263 and CVE-2026-34260: SAP Commerce Cloud unauthenticated code injection and S/4HANA Enterprise Search SQL injection (May 2026 Patch Day)

SAP released its May 2026 security notes on 2026-05-12, among them Note 3733064 for CVE-2026-34263 in SAP Commerce Cloud (HY_COM 2205, COM_CLOUD 2211 and 2211-JDK21) and Note 3724838 for CVE-2026-34260 in SAP S/4HANA's Enterprise Search for ABAP (SAP_BASIS 751 to 758 and 816), both rated Critical at CVSS 9.6 (SAP, May 2026); Onapsis counts seventeen new and updated notes, three of them HotNews (Onapsis, 2026-05-12). CVE-2026-34263 stems from improper Spring Security configuration that lets an unauthenticated user perform malicious input injection resulting in arbitrary server-side code execution (EUVD-2026-29372); Onapsis traces it to an overly permissive security configuration with improper rule ordering that allows malicious configuration upload and code injection (Onapsis, 2026-05-12), and SAP's vector requires user interaction (SAP, May 2026). CVE-2026-34260 lets an authenticated attacker inject SQL through user-controlled input that the application concatenates into its queries (EUVD-2026-29371); the affected code only allows reading data, so confidentiality and availability are affected but integrity is not (Onapsis, 2026-05-12). The cycle's third HotNews note, 3747787 of 2026-04-30, covers the variant of the Shai-Hulud worm that injected malicious code into SAP Cloud Application Programming npm packages available for two to four hours on 2026-04-29 (Onapsis, 2026-05-12). SAP makes no mention of exploitation in the wild (SecurityWeek, 2026-05-12).

The vulnerability is caused by an overly permissive security configuration with improper rule ordering, allowing an unauthenticated user to perform malicious configuration upload and code injection, resulting in arbitrary server-side code execution.

Fortunately, the affected source code only allows read access to data so that integrity is not impacted.

Onapsis 2026-05-12

SAP makes no mention of any of these vulnerabilities being exploited in the wild, but users are advised to apply the patches as soon as possible.

SecurityWeek 2026-05-12
Correctionrun 2026-10-21T1148Z-auditactionsaffected_productsbodyclassificationcvesevidenceheadlineprioritysourcessourcing_notesummarytechniquestitle

CVE-2026-34260 needs authentication: it is an SQL injection through user-controlled input in S/4HANA's Enterprise Search for ABAP (EUVD-2026-29371), and because the affected code only reads data its integrity impact is none, so the reach is database reads and availability rather than full read and write (Onapsis, 2026-05-12). CVE-2026-34263 in Commerce Cloud needs no authentication but SAP's vector requires user interaction (SAP, May 2026), and its root cause is improper Spring Security configuration (EUVD-2026-29372). HotNews Note 3747787 for the SAP Cloud Application Programming npm compromise was released on 2026-04-30 (Onapsis, 2026-05-12), and none of the cited sources links either product to Swiss federal programmes or says how often Commerce Cloud deployments face the internet.

vulnerability13 May 05:00Zmulti-sourceOpen finding →

03Research, reports & policy3 items

NOTABLEupdatedNATOB2

TrickMo C: Android banking trojan moves its C2 onto The Open Network and turns infected phones into SOCKS5 and SSH network pivots

ThreatFabric identified a new TrickMo variant between January and February 2026 that is progressively replacing its predecessor in campaigns against banking and wallet customers in France, Italy and Austria (ThreatFabric, 2026-05-11). The on-device feature set is largely unchanged: once the user grants the accessibility-service permission the bot coerces, the operator gets overlay credential phishing, keylogging, screen recording and streaming, remote control, and SMS and notification interception that silently suppresses one-time-password messages, while the offensive code arrives as a module fetched from operator infrastructure at runtime (ThreatFabric, 2026-05-11). Campaign tags point to TikTok-themed apps delivered over Facebook (ThreatFabric, 2026-05-11); the droppers pose as adult-friendly versions of TikTok and the malware as Google Play Services (The Hacker News, 2026-05-12).

Command and control now runs over The Open Network: an embedded native TON proxy starts on a loopback port, the bot's HTTP client sends every request through it to .adnl hostnames resolved inside the TON overlay, and the few clearnet names it still needs are resolved over public DNS-over-HTTPS (ThreatFabric, 2026-05-11). New operator commands run curl, DNS lookups, ping, multi-port TCP-connect probes and traceroute from the device's network position, open SSH local-forward and remote-forward tunnels, the latter reaching into the corporate or home LAN the phone is on, and start an authenticated SOCKS5 proxy whose traffic appears to come from the victim's IP and so defeats IP-based fraud checks (ThreatFabric, 2026-05-11). The Pine hooking framework and a full set of NFC permissions are bundled but unused, which ThreatFabric reads as preparation for features delivered at runtime later (ThreatFabric, 2026-05-11).

Traffic-pattern detection at the network edge sees only TON traffic, which is encrypted and indistinguishable from any other TON-enabled application's outbound flow.

these dormant components suggest the operators are preparing the platform for future runtime-delivered features

ThreatFabric 2026-05-11

Dropper apps containing the malware masquerade as adult-friendly versions of TikTok through Facebook

The Hacker News 2026-05-12
Improvementrun 2026-10-21T1148Z-auditaffected_productsbodyclassificationevidenceheadlinesourcessourcing_notesummarytechniquestitle

ThreatFabric's point about the TON transport is that traffic analysis at the network edge sees only encrypted TON traffic, indistinguishable from any other TON application's, and that domain takedowns lose effect because the endpoints live inside the overlay (ThreatFabric, 2026-05-11); it does not say TON shares standard internet routes, and no source describes "unusual entitlements" as a signal. The Pine hooking framework and the NFC permissions are bundled but not exercised, which ThreatFabric reads as capability reserved for features delivered at runtime later, possibly NFC-targeted, not as contactless-payment interception in development (ThreatFabric, 2026-05-11). The droppers are TikTok look-alikes delivered through Facebook campaigns (ThreatFabric, 2026-05-11; The Hacker News, 2026-05-12); no source says ads. The observable activity is on the device: accessibility-service grants to an app posing as Google Play Services, and network probes originating from a phone inside the LAN (ThreatFabric, 2026-05-11).

research13 May 05:00Zmulti-sourceOpen finding →
NOTABLEupdatedNATOA2

NCSC-UK: 10 questions to ask when using AI models to find vulnerabilities

NCSC-UK published on 2026-05-11 ten questions to answer before using an AI model to find vulnerabilities, written by its Head of Vulnerability Management Group (NCSC-UK, 2026-05-11). It warns that just finding vulnerabilities "does nothing to improve your security" and "could even make your security worse", and asks whether a process exists to receive, prioritise and fix what AI finds without overwhelming teams (NCSC-UK, 2026-05-11). Its risk questions cover information leakage, securing the infrastructure used, sandboxing the system so it can only talk to the LLM and the code base, whether to give it production access, the permissions given to the LLM, spending all the money, time and people on finding rather than fixing, terms and data-retention policies, and legality (NCSC-UK, 2026-05-11). It ties prioritisation to exploitability: of more than 40,000 CVEs assigned in 2025, CISA's catalogue tracked about 400 new ones as exploited, around 40 of them zero-days when first exploited (NCSC-UK, 2026-05-11).

just finding vulnerabilities does nothing to improve your security. You could even make your security worse.

When using AI to find vulnerabilities prioritise your external attack surface. You should also look for ways to verify the results, by using both AI and humans.

NCSC-UK 2026-05-11
Correctionrun 2026-10-21T1148Z-auditaffected_productsbodyclassificationevidenceheadlinesourcessummarytechniquestitle

The ten questions contain no item on false-positive overhead for defenders; the nearest is resourcing: "How can I avoid spending all my money/time/people finding vulnerabilities, and have nothing left to fix them?" (NCSC-UK, 2026-05-11). NCSC-UK's warning is that just finding vulnerabilities does nothing to improve security and could even make it worse (NCSC-UK, 2026-05-11); it does not say AI-accelerated discovery leaves an organisation worse off or criticise buying AI tools, and its data and permissions questions ask about information leakage, production access and what permissions the LLM is given, without mentioning secrets or a compromised agent.

research13 May 05:00Zsingle-source · national CERTOpen finding →
Sources: NCSC-UK
NOTABLEupdatedNATOB2

Microsoft MDASH: multi-model agentic scanning harness credited with 16 Windows networking and authentication CVEs, four of them Critical remote code execution flaws

Microsoft's Autonomous Code Security team disclosed MDASH on 2026-05-12, a harness that orchestrates more than 100 specialised agents across an ensemble of frontier and distilled models in a prepare, scan, validate, dedupe and prove pipeline: debater agents argue each finding's reachability and exploitability, and the prove stage constructs and executes triggering inputs where the bug class admits it, to prove the vulnerability exists (Microsoft, 2026-05-12). Microsoft lists 16 CVEs its engineering teams found with MDASH across the Windows network stack and adjacent services, 10 kernel-mode and 6 user-mode, most of them reachable from a network position with no credentials (Microsoft, 2026-05-12).

Four are Critical remote code execution flaws: CVE-2026-33827, a race-driven use-after-free in tcpip.sys; CVE-2026-33824, a double-free in the IKEEXT service, which runs as LocalSystem, triggered by two UDP packets; CVE-2026-41089, an unauthenticated CLDAP filter stack overflow in netlogon.dll; and CVE-2026-41096, a heap out-of-bounds flaw in dnsapi.dll triggered by a crafted UDP DNS response (Microsoft, 2026-05-12). Microsoft released the first two on 2026-04-14, scored 8.1 with high attack complexity and 9.8, both rated Exploitation Less Likely (MSRC, CVE-2026-33827; MSRC, CVE-2026-33824), and the other two on 2026-05-12, both scored 9.8, the Netlogon flaw rated Exploitation Less Likely and the DNS client flaw Exploitation Unlikely (MSRC, CVE-2026-41089; MSRC, CVE-2026-41096). The MDASH write-up describes the tcpip.sys trigger as crafted IPv4 packets carrying the Strict Source and Record Route option (Microsoft, 2026-05-12), while the advisory's FAQ describes a crafted IPv6 packet sent to a node where IPsec is enabled (MSRC, CVE-2026-33827).

Microsoft's benchmark figures are its own: all 21 vulnerabilities planted in a private test driver found with zero false positives, recall of 96% on 28 historical MSRC cases in clfs.sys and 100% on 7 in tcpip.sys, which it describes as retrospective recall on internal code with a finite case count, and an 88.45% success rate on CyberGym's 1,507 vulnerability reproduction tasks drawn from 188 OSS-Fuzz projects (Microsoft, 2026-05-12). The harness is used by Microsoft security engineering teams and tested by a small set of customers in a limited private preview (Microsoft, 2026-05-12; The Register, 2026-05-13).

Across the Windows network stack and adjacent services, today’s Patch Tuesday includes 16 CVEs our engineering teams found using codename MDASH.

Constructs and executes triggering inputs where the bug class admits it.

These vulnerabilities are 10 kernel-mode / 6 usermode.

a bare Start-Service IKEEXT with no responder policy is not vulnerable

Microsoft Security Blog 2026-05-12
Correctionrun 2026-10-21T1148Z-auditaffected_productsbodyclassificationevent_dateevidenceheadlinesourcessourcing_notesummarytechniquestitle

Microsoft describes MDASH's prove stage as constructing and executing triggering inputs where the bug class admits it, to prove that a vulnerability exists (Microsoft, 2026-05-12); it does not say every finding was proven exploitable, and its advisories rate CVE-2026-33827, CVE-2026-33824 and CVE-2026-41089 Exploitation Less Likely and CVE-2026-41096 Exploitation Unlikely (MSRC, CVE-2026-33827; MSRC, CVE-2026-33824; MSRC, CVE-2026-41089; MSRC, CVE-2026-41096). CVE-2026-33827 and CVE-2026-33824 were patched in April 2026, not in the May update, and six of the 16 bugs are user-mode while the cohort also covers http.sys and telnet.exe, so they are not all network-stack kernel flaws (Microsoft, 2026-05-12).

Microsoft did not call the 16 CVEs a first production run, and gives no June date: the harness is already being tested by a small set of customers in a limited private preview (Microsoft, 2026-05-12). The benchmark figures come from Microsoft's blog rather than The Register, and the tcpip.sys recall covers 7 historical MSRC cases (Microsoft, 2026-05-12). No cited source describes DirectAccess or Always-On VPN deployment in Swiss or EU government estates; the IKEv2 flaw is reachable only on hosts configured as IKEv2 responders (Microsoft, 2026-05-12).

research13 May 05:00Zmulti-sourceOpen finding →

04Updates to prior coverage2 items

HIGHCVE-2026-0300exploitedupdatedNATOA1

CVE-2026-0300: PAN-OS User-ID Authentication Portal unauthenticated root RCE, exploited by a likely state-sponsored cluster; fixed builds for every supported branch

First published 2026-05-12 · open finding →

Updaterun 2026-05-13-c148b9a5actionscvesevidencetagsbody

UPDATE (originally covered 2026-05-12): Palo Alto Networks released the first wave of patched PAN-OS builds on 2026-05-13 for the actively-exploited Captive Portal pre-auth RCE, covering PAN-OS 10.2, 11.1, 11.2 and 12.1 (Palo Alto Networks PSIRT, last updated 2026-05-07; patch table confirmed 2026-05-13).

Palo Alto Networks released the first wave of fixed PAN-OS builds on 2026-05-13, covering the 12.1, 11.2, 11.1 and 10.2 branches; the advisory still lists 12.1.7 with an ETA of 2026-05-28 (Palo Alto Networks PSIRT). Prisma Access, Cloud NGFW and Panorama remain unaffected, and Threat ID 510019 remains the interim control for unpatched devices on PAN-OS 11.1 or later (Palo Alto Networks PSIRT). CISA lists the flaw in its Known Exploited Vulnerabilities catalog (CISA KEV catalog, 2026-05-06).

NOTABLEupdatedNATOB2

Instructure Canvas breach: ShinyHunters exploits a Free-for-Teacher flaw twice, defaces about 330 login portals, and Instructure reaches an agreement with the actor

First published 2026-05-08 · open finding →

Updaterun 2026-05-13-c148b9a5regionssourcesbody

UPDATE (originally covered 2026-05-12): Late on 2026-05-11, US House Homeland Security Committee Chairman Andrew Garbarino sent a formal letter to Instructure CEO Steve Daly ahead of the 2026-05-12 ShinyHunters extortion deadline, demanding a briefing by 2026-05-21 on the circumstances of both Canvas intrusions …

Late on 2026-05-11, US House Homeland Security Committee Chairman Andrew Garbarino wrote to Instructure CEO Steve Daly requesting a briefing by 2026-05-21 on both Canvas intrusions, the volume of data accessed, containment measures and coordination with federal law enforcement and CISA (The Record, 2026-05-12); (The Register, 2026-05-12). The same evening Instructure announced its agreement with ShinyHunters (The Record, 2026-05-12). ShinyHunters' claim covers about 275 million records from about 8,800 schools (The Register, 2026-05-12).

05Deep dive1 item

NOTABLECVE-2026-45321exploitedupdatedNATOA1

Mini Shai-Hulud hits TanStack: a pull_request_target Pwn Request, Actions cache poisoning and OIDC token theft from runner memory published 84 malicious @tanstack versions with valid provenance

On 2026-05-11 between 19:20 and 19:26 UTC an attacker published 84 malicious versions across 42 @tanstack/* npm packages by chaining a pull_request_target "Pwn Request", GitHub Actions cache poisoning across the fork and base-repository trust boundary, and extraction of an OIDC token from the runner's memory; no npm token was stolen and the publish workflow itself was not modified (TanStack, 2026-05-11). StepSecurity, which detected the publishes, and Wiz attribute the wave to TeamPCP's Mini Shai-Hulud worm, which hit UiPath, Mistral AI and other npm packages the same day, though Wiz found a payload bug that leaves the malware non-functional in the @uipath and @mistralai packages (StepSecurity, 2026-05-11; Wiz, 2026-05-12). The worm's first wave had compromised four official SAP npm packages on 2026-04-29 through malicious preinstall scripts (The Register, 2026-04-30). The TanStack compromise is CVE-2026-45321, scored 9.6 (TanStack advisory, 2026-05-11), which CISA added to its Known Exploited Vulnerabilities catalog on 2026-05-27 with known ransomware campaign use (CISA KEV catalog, 2026-05-27).

The chain, from TanStack's timeline (TanStack, 2026-05-11):

  1. Staging. On 2026-05-10 the attacker forked TanStack/router into a repository renamed to evade fork-list searches and committed a roughly 30,000-line bundled JavaScript payload under the history package, with a fabricated author identity and a commit message that suppressed CI on push; the post-mortem names two attacker GitHub accounts.
  2. Pwn Request. On 2026-05-11 at about 10:49 a pull request titled as a work-in-progress build simplification triggered the bundle-size and labeler workflows, both on pull_request_target, which needs no first-time-contributor approval. At 11:11 the benchmark job checked out the PR's merge ref and ran pnpm install and the bundle-size build, executing the payload.
  3. Cache poisoning. The payload wrote into the pnpm store under the key the release workflow computes, and the cache action's post-job save stored the 1.1 GB poisoned store in the base repository's cache scope; the job's read-only permissions: did not stop it, because cache writes use a runner-internal token. At 11:31 the attacker force-pushed the PR back to the main branch head, leaving a zero-file change, and closed it.
  4. Detonation. At 19:15 a re-run of the release workflow, and at 19:16 a fresh run triggered by a legitimate merge, restored the poisoned store. Attacker binaries run during the build located the Runner.Worker process, read its memory through /proc/<pid>/maps and /proc/<pid>/mem, took the OIDC token the runner mints in memory for a job with id-token: write, and posted the packages directly to the npm registry, using the same memory-dump script as the tj-actions/changed-files compromise of March 2025. Tests failed, so the workflow's own Publish Packages step was skipped.

The packages carry valid SLSA provenance tied to TanStack's release workflow; provenance proves which pipeline built a package, not that the pipeline behaved, so a compromised build step yields a validly attested malicious package (StepSecurity, 2026-05-11).

The payload. Installing an affected version resolves an optional dependency on a github: URL naming the TanStack repository, which GitHub serves from an orphan commit in the fork network; npm installs that commit's dependencies, including the Bun runtime, and runs its prepare lifecycle script, which executes a 2.3 MB obfuscated JavaScript file shipped at the tarball root and then exits with an error so npm silently discards the optional install (TanStack advisory, 2026-05-11). Behind three obfuscation layers, it queries the AWS instance metadata service with the IMDSv2 flow and the ECS credential endpoint, reads the local Vault agent, reads files from more than 100 hardcoded paths, and in CI pipes a memory-reading script to sudo python3 to pull every value the runner marks as secret (StepSecurity, 2026-05-11); NCSC-CH groups the paths as cloud credentials, SSH, Git, developer and AI tools, crypto wallets, VPN configurations, messaging apps, shell history and Kubernetes (NCSC-CH, 2026-05-12). It exfiltrates encrypted data through the Session messenger's file-upload network and through GitHub commits on branches named to look like Dependabot's, spreads by finding npm tokens that bypass two-factor checks, enumerating the maintainer's other packages and, in CI, exchanging the OIDC token for per-package publish tokens, and writes workflows disguised as CodeQL analysis that serialise all repository secrets (StepSecurity, 2026-05-11). Persistence is a Claude Code SessionStart hook, a VS Code folder-open task, and a LaunchAgent on macOS or systemd user service on Linux that checks the GitHub API every 60 seconds and deletes the user's home directory once the stolen token is revoked (NCSC-CH, 2026-05-12).

No npm tokens were stolen and the npm publish workflow itself was not compromised.

Cache writes use a runner-internal token

TanStack 2026-05-11

Yet the packages were published during the same window.

A compromised build step can produce a validly-attested but malicious package.

Do not revoke these tokens before isolating and imaging the machine

StepSecurity 2026-05-11

On April 29, TeamPCP compromised four official npm packages from the SAP JavaScript and cloud application development ecosystem and published the poisoned releases between 09:55 and 12:14 UTC.

The Register 2026-04-30
Correctionrun 2026-10-21T1148Z-auditaffected_productsbodyclassificationcvesevent_dateevidenceheadlinesectorssourcessourcing_notesummarytagstechniquestitle

Mini Shai-Hulud's first wave hit four official SAP npm packages on 2026-04-29 with malicious preinstall scripts (The Register, 2026-04-30). In the TanStack compromise no reconnaissance step is reported; the pull request added a bundled payload file that the pull_request_target benchmark job executed when it checked out and built the PR's merge ref, rather than slipping in lockfile, dependency or .github changes, and the release job never reached its publish step: tests failed, Publish Packages was skipped, and the malware published directly with the OIDC token it read from runner memory (TanStack, 2026-05-11). No npm token was needed for that publish, but the payload itself steals credentials and spreads with them (StepSecurity, 2026-05-11).

The SLSA provenance on the malicious packages is genuine, an attestation of a compromised pipeline rather than a forgery, and no source ties it to EU procurement (StepSecurity, 2026-05-11). The payload runs from the prepare script of an optional git dependency, not from postinstall (TanStack advisory, 2026-05-11), and its memory scraper runs as sudo python3, so a filter for non-root readers of /proc/<pid>/mem misses it (StepSecurity, 2026-05-11). The compromise is CVE-2026-45321, scored 9.6 (TanStack advisory, 2026-05-11) and listed by CISA as known exploited on 2026-05-27 (CISA KEV catalog, 2026-05-27). The cache action has no save: false input; restore-only caching is the separate restore action (actions/cache README), npm verifies attestations with npm audit signatures (npm Docs), and GitHub Security Lab warns that combining the pull_request_target trigger with an explicit checkout of an untrusted pull request may lead to repository compromise (GitHub Security Lab, 2021-08-03).

threat13 May 05:00Zmulti-sourceOpen finding →

06Action items2 items

Verification & coverage notes1 run

2026-05-13-c148b9a5 · Claude Opus 4.7 · 15 entries published

  • Dropped items / coverage gaps:
    • CVE-2026-41901, Thymeleaf SSTI sandbox bypass. Initially surfaced by S2 (research returning ENISA EUVD-2026-29872 hit and CSO Online corroborating link) and composed into § 2. Phase 5.7 verifier flagged the source dates: the GitHub Security Advisory GHSA-c9ph-gxww-7744 is dated 2026-04-29 (not 2026-05-12), placing the disclosure 14 days outside the 36-hour recency window; the CSO Online article cited as corroborating is dated 2026-04-17 and concerns a different Thymeleaf CVE (CVE-2026-40478). Per PD-7 the item is dropped from § 2; the ENISA EUVD claim could not be independently re-verified in this iteration (EUVD direct fetch is SPA-only and the bridge fetcher hit an SSL cert-validity error during the verification pass). Operators with Spring Boot Java applications should patch Thymeleaf to 3.1.5.RELEASE in line with the April GHSA. Source: GitHub Security Advisory GHSA-c9ph-gxww-7744, 2026-04-29 · Tags: vulnerabilities, rce, patch-available · Region: global · CVE: CVE-2026-41901 · CVSS: 9.0 · Vector: user-interaction · Auth: pre-auth · Status: patch-available
    • Odido (Netherlands) compensation refusal, Dutch DPA / criminal investigation, CUIC class action (NL Times, 2026-05-12). Surfaced by S4 but the primary URL https://nltimes.nl/2026/05/12/odido-rules-compensation-massive-cyberattack-affecting-62-million-accounts is behind Cloudflare's Managed Challenge and could not be fetched in-run by either the routine UA or the bridge fetcher (nltimes.nl is not in the bridge allow-list). Search-engine snippets confirmed the publication date, headline and key facts (CEO Søren Abildgaard, ShinyHunters vishing of Salesforce, 6.2M accounts, 350k registered for CUIC class action, Dutch Public Prosecution Service criminal investigation, AP / ILT investigations), but PD-2 requires that every cited URL be one the agent actually fetched. Dropped from § 1 rather than carry an unverifiable primary; will re-attempt next run via WebSearch fallback content or alternative outlet, Source: Techzine, 2026-02-16 · The Register, 2026-02-27 · Tags: data-breach, identity · Region: europe · Sector: telco
    • Kaspersky State of Ransomware 2026 (Securelist, 2026-05-12). PD-9 annual report. Surfaced by S3 with novel headline (PE32 ransomware using ML-KEM/Kyber1024 post-quantum KEM; encryptionless extortion now dominant; Qilin displaces RansomHub; RDWeb portal targeting overtakes phishing). Not promoted to § 3 / § 5 in this brief, yesterday's brief (2026-05-12) already absorbed the annual-report deep-dive slot (GTIG AI Threat Tracker) and PD-3 demotes the same category one rank when in the last-7-day window. Will pick up the PE32 / post-quantum element as a § 3 research item in a later brief if it gains corroborating independent analysis, Source: Securelist (Kaspersky), 2026-05-12 · Tags: ransomware, supply-chain · Region: global · Sector: public-sector
    • CVEs from Microsoft May Patch Tuesday not surfaced in § 2. ~30 Critical CVEs landed in this cycle; § 2 cherry-picked the four most operationally significant (Netlogon, DNS Client, SSO Plugin, Dynamics 365) plus the four Word Preview Pane RCEs. The remaining MDASH-discovered network-stack CVEs are referenced indirectly in § 3 (MDASH article). Operators should review the full Tenable / ZDI breakdowns for their environment.
  • Single-source items:
    • NCSC-UK "10 questions to ask when using AI models to find vulnerabilities", flagged [SINGLE-SOURCE] in § 3; national-CERT carve-out applies (NCSC-UK is primary disclosing party for its own guidance).
  • Contradictions:
    • CVE-2026-26083 (FortiSandbox) CVSS: NCSC-CH cites 9.1 per Fortinet's own PSIRT, NVD's initial assignment is 9.8. Both indicate Critical; brief lists both ("9.1 (NCSC-CH per the vendor advisory) and 9.8 (NVD's initial assignment)"). Operators should treat as critical pending convergence; per-CVE breakdown carried in § 2 footer.
    • Microsoft May Patch Tuesday CVE count: Tenable reports 118, BleepingComputer reports 120, ZDI / The Register reports up to 138 depending on whether developer-tools and Azure-only items are included. Brief uses "120+" as a conservative summary.
    • Canvas affected-institution count: The Register cites ~8,800 colleges, universities and K-12 schools; The Record cites ~9,000. Brief uses 8,800 (per The Register, the primary cited) with The Record's figure surfaced inline.
  • Reduced-confidence items: None this run beyond the dropped Odido entry above.
  • Verification iteration 1 disposition (Opus): Eight F3/F4 truth findings applied as remediations, corrected NCSC-UK date (2026-05-12 → 2026-05-11), corrected Canvas narrative (Garbarino letter 2026-05-11 ahead of 2026-05-12 deadline), corrected Foxconn site-count phrasing, removed unsourced Nitrogen initial-access TTP chain, removed unverifiable ENISA EUVD claims on Exim and Thymeleaf, dropped Thymeleaf as out-of-window. Two F11 advisory findings (Foxconn TTP-chain advisory paired with the F4 fix; Canvas-institution-count) addressed.
  • Verification iteration 2 disposition (Sonnet rotation): Five findings (1 truth, 3 editorial, 1 advisory), (a) F1 broken-URL flag on https://fortiguard.fortinet.com/psirt/FG-IR-26-128: fresh re-fetch in this iteration returned 200 with the full advisory body (CVE-2026-44277, CVSS 9.1 per vendor, fixed in 6.5.7 / 6.6.9 / 8.0.3+, internal-audit discovery); the verifier's 403 was a transient UA-filter on the host; no source-list change required. (b) F3 Centreon wording corrected: "command injection (effectively RCE in Centreon MBI)" instead of "RCE"; "XSS (Centreon Map, CVSS 6.8)" instead of "reflected / stored XSS". (c) F5 BWH Switzerland claim removed (sources do not call out Swiss properties specifically). (d) F10 Centreon 24.04.x branch added to affected versions (MBI only per the vendor bulletin). (e) F11 advisory ATT&CK mapping in § 5 step 8 corrected from T1647 Plist File Modification to T1553.002 Subvert Trust Controls: Code Signing, with an explicit note that no current sub-technique precisely maps to SLSA-L3 provenance forgery via OIDC abuse.
  • Verification iteration 3 disposition (Opus rotation): Four findings (1 truth, 1 editorial, 2 advisory) at truth + editorial = 2 and no F1/F4, early-exit threshold reached per v2.50. Remediations applied before publish: (a) F3 CVE-2026-44277 / CVE-2026-26083 CVSS converged on vendor PSIRT canonical value 9.1 throughout § 0 / § 2 / § 6 (iter-2 had left 9.8 in the body even after § 7 noted the discrepancy); CVSS table footnote re-worded; (b) F11a Exim § 6 Action Item Tags / Status enisa-critical flag removed (iter-1 F4 fix propagated to § 6); (c) F5 BWH country list (Germany, France, Italy) replaced with "multiple EEA jurisdictions", neither cited source enumerates per-country exposure; (d) F11b advisory: Canvas UPDATE removed the "via support-ticket access" specifier; neither re-fetched source confirms that detail. Final iteration verdict: NEEDS_FIXES (truth=1 + editorial=1 = 2 ≤ early-exit threshold). Three iterations, model-rotated (Opus / Sonnet / Opus). Brief publishes with the four iter-3 remediations applied; no residual findings carried beyond what § 7 already documents.
  • Sub-agents: S1, S2, S3, S4 all returned within budget. No stalled sub-agents.
  • Fetch failures / bridges used: tools/fetch_source.py bridge used by S1, S2, S3, S4 for ncsc-csh recent / post (NCSC-CH SPA), cisa-kev, enisa-euvd recent, bsi-rss, and url for CERT-FR / BleepingComputer / Centreon. databreaches-net, inside-it-ch and parts of bleepingcomputer returned 403 (Cloudflare Managed Challenge) on every UA per the documented host behaviour, WebSearch fallback used. advisories-ncsc-nl returned 404 on the speculative ID (NCSC-NL CSAF advisory ID guessing not viable without a fresh index); coverage gap for that source this run. ico-uk sitemap fetched but no in-window enforcement actions found; freshest enforcement (South Staffordshire) already covered 2026-05-12. ENISA EUVD direct fetch hit SSL cert-validity error in the verification iteration, bridge worked for the sub-agents but not at verifier time.
  • Coverage gaps: databreaches-net (Cloudflare-blocked, WebSearch fallback only); inside-it-ch (Cloudflare-blocked, WebSearch fallback only); advisories-ncsc-nl (CSAF advisory-ID enumeration failed, no in-window items surfaced); ico-uk (no in-window enforcement); nltimes.nl (Cloudflare-blocked, dropped Odido item logged above).

Unmatched action items (migrated)

  • Patch Fortinet FortiAuthenticator and FortiSandbox now. Pre-auth RCEs in two appliance classes that anchor public-sector identity and SOC pipelines. Update FortiAuthenticator to 6.5.7 / 6.6.9 / 8.0.3; update FortiSandbox to 4.4.9 / 5.0.2 / Cloud 5.0.6. Cloud 23 and 24 require migration, not in-place patching. Even after patching, restrict the management interface to admin / SOC source IPs at the perimeter; Fortinet's PSIRT explicitly notes this as the residual hardening. See § 2
  • Audit npm / pnpm lockfiles for Mini Shai-Hulud impact. Search every CI/CD pipeline and developer workstation for malicious versions across @tanstack/*, @uipath/*, @mistralai/*, @opensearch-project/opensearch, @squawk/*, @draftlab/*, @tallyui/* per the StepSecurity / Wiz manifests. UiPath impact is the highest-priority public-sector signal. Before revoking any GitHub PAT or npm token, sanitise the developer machine first, the worm's gh-token-monitor triggers rm -rf ~/ on revocation. Pin all pull_request_target workflows to SHA-locked action versions; gate fork-reachable workflows away from actions/cache writes. See § 4 / § 5
  • Patch Exim on every Debian / Ubuntu mail relay. Run exim -bV | grep GnuTLS on each MTA, if present, upgrade to Exim 4.99.3. Interim workaround if patching is delayed: set CHUNKING_ADVERTISE_HOSTS = (empty) in exim4.conf to suppress BDAT. Expect public exploit-tooling within days; XBOW's disclosure includes full chain traces. See § 2
  • Upgrade SPIP to 4.4.14 and apply the Centreon April 2026 monthly bulletin. SPIP RCE affects French and francophone Swiss-canton CMS deployments, gate ecrire/ to a known admin source set at the reverse proxy. Centreon update queue: any NOC running 24.10.x / 25.10.x Infra Monitoring with Anomaly Detection / Auto Discovery / AWIE / BAM / DSM / License Manager / MAP / MBI / Open Tickets. See § 2
  • Hunt for Foxconn-Nitrogen-style precursor patterns on Windows servers and ESXi. Detection concepts in § 1: unsigned installers spawning cmd.exe / powershell.exe -enc from %TEMP% / %APPDATA% on Windows servers; vmkfstools / esxcli invocations from non-administrator sessions on ESXi /var/log/shell.log. Treat the Nitrogen ESXi decryptor bug as a strategic backup-integrity test; hypervisor-layer recovery from Nitrogen is mathematically impossible regardless of ransom posture. See § 1

Migrated from briefs/2026-05-13.md (v2).