CTIPilot
AI-generated · no human review · verify critical claims against the linked source. how it works →
‹Sat · 09 May 2026›
All daily briefs →
Daily brief · UTC day

Saturday, 9 May 2026

11 verified findings from 1 run · 3 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. 01An unauthenticated Perl eval() flaw and five more in SEPPmail's gateway, with fixes shipped from 15.0.2.1 to 15.0.4. NCSC-CH's advisory of 2026-05-08 lists six CVEs (CVSS 6.9 to 9.3) in the SEPPmail Secure E-Mail Gateway as fixed in 15.0.4, the release that removed development test endpoints left active in the beta GINA v2 interface, with the unauthenticated code execution flaw CVE-2026-44128 (9.3) already fixed in 15.0.2.1. InfoGuard's write-up traces that flaw to Perl eval(), places the insecure deserialization fix (CVE-2026-44126) in 15.0.3, and adds the LFT path traversal CVE-2026-2743 (10.0 per The Hacker News). No exploitation is reported. →
  2. 02LiteLLM SQL injection: attackers targeted stored credentials 36 hours after disclosure; patching does not undo exposure. CVE-2026-42208 (CVSS 4.0 9.3) is a SQL injection in LiteLLM Proxy's API-key verification, reachable without credentials from any LLM API route in versions 1.81.16 to 1.83.6 and fixed in 1.83.7. Sysdig saw the first exploitation attempt 36 hours after the advisory, with queries against the tables holding virtual keys, stored provider credentials and proxy configuration, but no follow-on use of extracted keys; CISA added it to its KEV catalog on 2026-05-08. →
  3. 03Dirty Frag: two kernel page-cache write flaws give local root; Microsoft saw su escalation that may be it or Copy Fail. Dirty Frag is two Linux kernel flaws, CVE-2026-43284 in the xfrm-ESP receive path and CVE-2026-43500 in RxRPC, that let a local user write into the page cache of read-only files and gain root deterministically. A public exploit appeared on 2026-05-07, mainline and distribution fixes are released, and Microsoft reports limited su-based escalation in the wild that may be Dirty Frag or Copy Fail. →

01Active threats, incidents & disclosures3 items

ROUTINEupdatedNATOB2

Berlin court orders Apobank to refund EUR 218,000 lost to a letter, QR-code and phone scam that enrolled the attackers' own device

Landgericht Berlin II ruled on 2026-04-22 (38 O 293/25) that Apobank may not refuse to refund about EUR 218,000 on grounds of gross negligence: a bank-branded letter asked the customer to register a new security procedure, its QR code led to a cloned banking portal, a caller posing as the bank guided her through the process, and photographing a QR code let the attackers enrol their own device; the judgment is not yet final and the customer bears only the EUR 50 statutory share (Borncity, 2026-05-09). Beyond the holding, the chamber remarked that the login and the attackers' new-device registration came from entirely different IP addresses and providers and that the bank did not block this discrepancy automatically (heise online, 2026-05-08).

Contradiction: Borncity has a QR code in the letter leading to the manipulated page (Borncity, 2026-05-09), while heise reports that the customer was asked to set up two-factor authentication during her usual login through her browser's favourites bar (heise online, 2026-05-08).

Correctionrun 2026-10-11T1307Z-auditkindtitleheadlinesummaryevent_dateprioritytagsregionssectorstechniquessourcesclassificationbody

The court's holding is that the customer was not grossly negligent, so the bank must refund the loss (Borncity, 2026-05-09); its comments on the IP and provider mismatch were remarks beyond the decision, which the claimant's lawyer called technically still amateurish (heise online, 2026-05-08). No cited source mentions PSD2 or a duty to trigger strong customer authentication, and the judgment is described as not yet final rather than under appeal.

incident09 May 05:00Zmulti-sourceOpen finding →
NOTABLEupdatedNATOB2

DAEMON Tools Lite supply-chain compromise: signed installers deliver a profiler worldwide, with a backdoor and QUIC RAT reserved for a dozen targets in Russia, Belarus and Thailand

Kaspersky found that DAEMON Tools installers distributed from the product's legitimate website and signed with the developer's certificates had been trojanised from 2026-04-08: in versions 12.5.0.2421 to 12.5.0.2434 the helper binaries DTHelper.exe, DiscSoftBusServiceLite.exe and DTShellHlp.exe, signed by AVB Disc Soft, were compromised, and version 12.6.0.2445 no longer contains the malicious behaviour (Kaspersky Securelist, 2026-05-05). At startup the trojanised binaries activate a backdoor that sends GET requests to a domain typosquatting the vendor's download site; when the server replies with a shell command, it runs through cmd.exe, which in the observed cases ran PowerShell to download an executable into C:\Windows\Temp and execute it, the first payload being a .NET information collector (Kaspersky Securelist, 2026-05-05). Kaspersky saw several thousand infection attempts in more than 100 countries, 10% of the affected systems belonging to organizations, and most received only the information collector (Kaspersky Securelist, 2026-05-05). A minimalistic backdoor reached about a dozen machines of government, scientific, manufacturing and retail organizations in Russia, Belarus and Thailand, and was used to deploy QUIC RAT against one organization, an educational institution in Russia; QUIC RAT supports HTTP, UDP, TCP, WSS, QUIC, DNS and HTTP/3 for command and control and can inject payloads into notepad.exe and conhost.exe (Kaspersky Securelist, 2026-05-05). Chinese strings in the collector may imply a Chinese-speaking actor, but Kaspersky does not attribute the compromise, and whether the intent is espionage or big-game hunting is unclear (Kaspersky Securelist, 2026-05-05). The developer says only the free Lite edition was affected (The Record, 2026-05-06).

Correctionrun 2026-10-11T1307Z-audittitleheadlinesummaryevent_dateprioritytagsregionssectorsaffected_productstechniquessourcesclassificationbody

Kaspersky reports no EU government victim: EU countries are among the more than 100 countries with infection attempts, which mostly delivered an information collector, while the backdoor reached about a dozen machines in Russia, Belarus and Thailand and QUIC RAT a single educational institution in Russia (Kaspersky Securelist, 2026-05-05). Its QUIC RAT description is limited to the C2 protocols and injection into notepad.exe and conhost.exe; the keylogging, SOCKS proxy and tunnelling capabilities given earlier had no source. Kaspersky does not attribute the compromise and calls its intent unclear.

threat09 May 05:00Zmulti-sourceOpen finding →
ROUTINEupdatedNATOB3

Zara customer emails exposed through a former Inditex technology provider; ShinyHunters claims Anodot token abuse and leaks the data

Inditex says databases hosted by a former technology provider, holding information about business relationships with customers in different markets, were accessed, and that customers' names, phone numbers, addresses, credentials and payment information were not; it has notified the relevant authorities but has not named the provider or attributed the breach (BleepingComputer, 2026-05-08). Have I Been Pwned found about 197,000 unique email addresses with product SKUs, order IDs and the market each support ticket originated in (Have I Been Pwned). ShinyHunters claims it stole the data with Anodot authentication tokens used against BigQuery and leaked a 140 GB archive (BleepingComputer, 2026-05-08); Have I Been Pwned describes the group's publication as a terabyte allegedly including 95 million support-ticket records (Have I Been Pwned).

Correctionrun 2026-10-11T1307Z-audittitleheadlinesummaryprioritytagsregionstechniquessourcesverificationclassificationbody

Inditex confirms only that a former technology provider's databases were accessed (BleepingComputer, 2026-05-08); the analytics-platform route is ShinyHunters' claim about Anodot tokens (BleepingComputer, 2026-05-08). The email, order and support-ticket fields come from Have I Been Pwned's analysis, and the records concern customers in different markets, and no source limits them to the EU (Have I Been Pwned) (BleepingComputer, 2026-05-08). The size of the leak differs between sources, 140 GB (BleepingComputer, 2026-05-08) against about a terabyte (Have I Been Pwned).

incident09 May 05:00Zmulti-sourceOpen finding →
HIGHCVE-2026-44128 +6updatedNATOA2

SEPPmail Secure E-Mail Gateway: unauthenticated code execution and five more flaws in NCSC-CH's advisory (CVE-2026-44128 et al., CVSS up to 9.3), fixed by 15.0.4

NCSC-CH's Security Hub post 12551 of 2026-05-08 lists six CVEs in the SEPPmail Secure E-Mail Gateway, scored from 6.9 to 9.3, with exploitation status unknown and vendor patches available since 2026-04-24 (NCSC-CH, 2026-05-08). SEPPmail's 15.0.4 release notes say test endpoints intended only for development had been left active in the beta GINA v2 web interface and are now removed (SEPPmail release notes). CVE-2026-44128 (9.3) is unauthenticated remote code execution through a parameter passed to Perl eval() and was already fixed in 15.0.2.1 (NCSC-CH, 2026-05-08) (InfoGuard Labs, 2026-05-18); CVE-2026-44126 (9.2), an insecure deserialization, was fixed in 15.0.3 (InfoGuard Labs, 2026-05-18), while NCSC-CH's advisory says the vulnerabilities were fixed in 15.0.4 (NCSC-CH, 2026-05-08); CVE-2026-44125 (9.3, missing authorization on GINA v2 endpoints), CVE-2026-44127 (8.8, file inclusion and arbitrary file deletion), CVE-2026-44129 (8.3, server-side template injection) and CVE-2026-7864 (6.9, exposure of sensitive information) were fixed in 15.0.4 (InfoGuard Labs, 2026-05-18); the scores are NCSC-CH's (NCSC-CH, 2026-05-08). The appliance is OpenBSD-based with Perl web applications, and self-registration, enabled by default, gives an attacker an authenticated session (InfoGuard Labs, 2026-05-18).

Updaterun 2026-05-20-a0f7b07factionscvesregionssectorssourcestagsbody

InfoGuard Labs, which found CVE-2026-2743 and the GINA v2 flaws, published its technical write-up on 2026-05-18 (InfoGuard Labs, 2026-05-18). CVE-2026-2743 (CVSS 10.0 (The Hacker News, 2026-05-19)) is a pre-authentication path traversal in the Large File Transfer (LFT) component's /v1/file.app endpoint that writes files as the nobody user; because nobody can write /etc/syslog.conf, and newsyslog runs every 15 minutes and makes syslogd reload its configuration, the write becomes command execution (InfoGuard Labs, 2026-05-18). Only instances with LFT, which InfoGuard says most customers enable, are affected; /v1/file.app returning HTTP 404 means not affected (InfoGuard Labs, 2026-05-18). CVE-2026-2743 is not new: SEPPmail's 15.0.3 release notes already close it (SEPPmail release notes). The Hacker News reported the write-up (The Hacker News, 2026-05-19).

Contradiction: SEPPmail's release notes place the CVE-2026-2743 fix in 15.0.3 (SEPPmail release notes), while InfoGuard's table lists 15.0.4 (InfoGuard Labs, 2026-05-18); 15.0.4 or later covers both readings. InfoGuard also places the CVE-2026-44126 fix in 15.0.3 (InfoGuard Labs, 2026-05-18), while NCSC-CH's advisory says the vulnerabilities were fixed in 15.0.4 (NCSC-CH, 2026-05-08), so 15.0.4 is the safe floor for that flaw too.

Correctionrun 2026-10-11T1307Z-audittitleheadlinesummarysectorstagscvesaffected_productstechniquessourcesverificationclassificationactionsbody

The fixes arrived across three releases: CVE-2026-44128 in 15.0.2.1, CVE-2026-44126 in 15.0.3 and the remaining four GINA v2 flaws in 15.0.4; 15.0.4.1 is an unrelated hotfix (InfoGuard Labs, 2026-05-18) (SEPPmail release notes). CVE-2026-2743 was closed in SEPPmail's 15.0.3 release, so it was not new in May; InfoGuard's May write-up gives 15.0.4 as its fix version (SEPPmail release notes) (InfoGuard Labs, 2026-05-18). CVE-2026-7864 exposes sensitive information and CVE-2026-44127 allows file inclusion and deletion; neither is code execution (InfoGuard Labs, 2026-05-18). SEPPmail AG is based in Neuenhof, not Steinach (SEPPmail).

vulnerability09 May 05:00Zmulti-sourceOpen finding →
HIGHCVE-2026-43284 +1updatedNATOB2

CVE-2026-43284 / CVE-2026-43500 Linux Dirty Frag: deterministic page-cache write to root through xfrm-ESP and RxRPC, public exploit, fixes released

Researcher Hyunwoo Kim (V4bel) published "Dirty Frag" on 2026-05-07, after the embargo had been broken by external factors and before patches or CVE ids existed (V4bel, 2026-05-07). It is two page-cache write flaws: CVE-2026-43284 in the xfrm-ESP (IPsec) receive path, in the kernel since 2017 and fixed in mainline commit f4c50a4034e6, and CVE-2026-43500 in RxRPC, in the kernel since 2023 and fixed in mainline commit aa54b1d27fe0 (V4bel, 2026-05-07). In both, a page-cache page planted with splice() sits in a fragment of a network buffer and the receiving code decrypts it in place, writing into the cached copy of a file: the ESP variant stores 4 attacker-chosen bytes per trigger into a setuid binary such as su, and the RxRPC variant stores 8 bytes whose value the attacker brute-forces through the session key, aimed at /etc/passwd; the result is deterministic root without a race window (V4bel write-up, 2026-05-07). The ESP variant needs the privilege to create a user namespace, which Ubuntu sometimes blocks through AppArmor; the RxRPC variant needs no namespace but depends on the rxrpc module, which most distributions do not ship and Ubuntu loads by default (V4bel, 2026-05-07). The kernel.org CNA scores CVE-2026-43284 CVSS 8.8, and NVD and CISA score CVE-2026-43500 7.8 (Ubuntu, 2026-05-08). Microsoft reports limited in-the-wild privilege escalation involving su that "may be indicative of techniques associated with either" Dirty Frag or Copy Fail: an SSH session ran an ELF binary that escalated through su, then edited a GLPI LDAP authentication file and deleted and read web session files (Microsoft Security Blog, 2026-05-08). Copy Fail is the earlier CVE-2026-31431 (Microsoft Security Blog, 2026-05-08).

The researcher tested the exploit on Ubuntu 24.04.4, RHEL 10.1, openSUSE Tumbleweed, CentOS Stream 10, AlmaLinux 10 and Fedora 44 (V4bel, 2026-05-07). Ubuntu publishes fixed kernels for every supported release from 18.04 to 26.04 (Ubuntu, 2026-05-08), and Red Hat states that fixes are released for all affected products and that CVE-2026-43500 does not affect Red Hat products (Red Hat RHSB-2026-003). Where a fix cannot be applied, NCSC-CH advises blocking or removing the esp4, esp6 and rxrpc modules (NCSC-CH advisory 12547, 2026-05-08), which affects IPsec and AFS where they are in use (Ubuntu, 2026-05-08); on RHEL, user.max_user_namespaces=0 blocks the ESP path (Red Hat RHSB-2026-003).

Updaterun 2026-05-11-migratedcvesevidenceregionssourcesbody

Microsoft's Defender Security Research Team reported on 2026-05-08 "limited in-the-wild activity where privilege escalation involving ‘su’ is observed, and which may be indicative of techniques associated with either “Dirty Frag” or “Copy Fail”": an external connection gained SSH access, spawned a shell and ran an ELF binary that immediately escalated through su (Microsoft Security Blog, 2026-05-08). Microsoft does not say which flaw the binary used.

Red Hat's bulletin RHSB-2026-003 lists the Dirty Frag CVEs, states that CVE-2026-43500 does not affect Red Hat products, and gives user.max_user_namespaces=0 as a targeted mitigation alongside blocking the ESP modules (Red Hat RHSB-2026-003). NCSC.ch's Security Hub post 12547 of 2026-05-08 noted a public proof of concept and advised temporarily blacklisting esp4, esp6 and rxrpc (NCSC-CH advisory 12547, 2026-05-08), and Belgium's CCB issued a parallel advisory (CCB Belgium, 2026-05-08).

The mainline fixes are f4c50a4034e6 for the ESP flaw and aa54b1d27fe0 for the RxRPC flaw (V4bel, 2026-05-07). Microsoft warns that modifications made before mitigation may persist in cached file content even after the vulnerable modules are disabled (Microsoft Security Blog, 2026-05-08). On Ubuntu systems where AppArmor blocks unprivileged user namespaces, the ESP path cannot be triggered, while the RxRPC path needs no namespace (V4bel, 2026-05-07). Blocking rxrpc breaks AFS and other RxRPC uses, so for those hosts the kernel update is the fix (Ubuntu, 2026-05-08).

Updaterun 2026-09-05T0409Z-intelupdated_atcvesentitiesclassificationtechniquessourcesactionssourcing_notebody

A related flaw, CVE-2026-46300 ("Fragnesia"), reopens this vulnerability's underlying page-cache-write primitive even on hosts already patched against CVE-2026-43284: a bug from 2013 in the code that merges packet fragments drops the marker that flags a fragment as page-cache-backed, which the original xfrm-ESP fix depends on to decide whether it is safe to decrypt in place (Aikido Security). Red Hat's own security bulletin groups all three CVEs under the collective "Dirty Frag" name, notes that the family is similar enough to the earlier Copy Fail flaw to be called "Copy Fail 2", and confirms CVE-2026-46300 affects supported Red Hat Enterprise Linux kernels; administrators who patched only against CVE-2026-43284/CVE-2026-43500 should verify the CVE-2026-46300 fix is applied too (Red Hat RHSB-2026-003). Public proof-of-concept exploits for this family now target Kubernetes specifically, extending the exposure to container-shared-kernel environments beyond the bare-metal/VM case originally described (Aikido Security).

Correctionrun 2026-10-11T1307Z-audittitleheadlinesummaryevent_datetagscvesaffected_productssourcesevidencesourcing_notebodyactions

Microsoft's observation of su-based escalation "may be indicative of techniques associated with either" Dirty Frag or Copy Fail (Microsoft Security Blog, 2026-05-08), so exploitation of these two CVEs is not confirmed. The flaws work by decrypting a splice()-planted page-cache page in place, not through DMA remapping (V4bel write-up, 2026-05-07); mainline fixed the ESP flaw in f4c50a4034e6 and the RxRPC flaw in aa54b1d27fe0, and distribution kernels are available (V4bel, 2026-05-07) (Ubuntu, 2026-05-08). Red Hat says CVE-2026-43500 does not affect its products and gives user.max_user_namespaces=0 as its targeted mitigation (Red Hat RHSB-2026-003). The write modifies the page cache in RAM (V4bel write-up, 2026-05-07) while the file on disk never changes (Aikido Security, 2026-09-04), so a file-write watch does not fire. The ESP variant stores 4 attacker-chosen bytes per trigger into a setuid binary such as su; the RxRPC variant stores 8 bytes whose value the attacker brute-forces through the session key, aimed at /etc/passwd (V4bel write-up, 2026-05-07).

Improvementrun 2026-10-21T1149Z-auditcvessourcessourcing_notebodyheadlinereferences

FIRST's EPSS probability was 0.0262 for CVE-2026-43500 and 0.9419 for CVE-2026-43284 on 2026-10-20 (FIRST.org EPSS API, 2026-10-20); on 2026-09-04 the two scores were 0.9286 and 0.9324 (FIRST.org EPSS API, 2026-09-04). Microsoft's report of 2026-05-08 is by its Defender Security Research Team (Microsoft Security Blog, 2026-05-08).

Builds on: Copy Fail turns an unprivileged Linux shell into root with one portable script; CISA lists it…

vulnerability09 May 05:00Zmulti-sourceOpen finding →
HIGHCVE-2026-42208exploitedupdatedNATOA2

CVE-2026-42208: LiteLLM Proxy pre-authentication SQL injection in API-key verification, first exploitation attempt 36 hours after the advisory, CISA KEV-listed (CVSS 4.0 9.3)

LiteLLM's advisory GHSA-r75f-5x8p-qvmc, published on 2026-04-20, describes a SQL injection in the proxy's API-key verification: the caller-supplied value was not passed to the database as a separate parameter, and the flaw is fixed in 1.83.7 (LiteLLM, 2026-04-20). Bishop Fox confirmed that versions 1.81.16 through 1.83.6 are exploitable without credentials from any LLM API route, through the Authorization: Bearer value reaching PrismaClient.get_data(); the response is always the same HTTP 401, so exploitation runs over a time-based blind channel using PostgreSQL's pg_sleep() (Bishop Fox, 2026-05-06). On a stock Docker Compose deployment with the Postgres sidecar, the application database user is a superuser, which raises the impact beyond reading metadata (Bishop Fox, 2026-05-06). Virtual keys are stored as SHA-256 hashes, but the provider credentials are stored for use (Bishop Fox, 2026-05-06) (Sysdig, 2026-04-27).

Sysdig observed the first exploitation attempt 36 hours and seven minutes after the advisory reached the global database, with queries against LiteLLM_VerificationToken (virtual keys, including the master key), litellm_credentials (stored upstream provider credentials) and litellm_config (Sysdig, 2026-04-27). Sysdig saw no follow-through: no authenticated calls with exfiltrated keys, no virtual-key minting through /key/generate and no reuse of provider credentials, so its finding is the speed and precision of the enumeration rather than a confirmed compromise; it still advises treating any internet-facing instance on a vulnerable version as compromised (Sysdig, 2026-04-27). CISA added the flaw to its Known Exploited Vulnerabilities catalog on 2026-05-08 (CISA KEV catalog, 2026-05-08). The published workaround, general_settings.disable_error_logs: true, closes only one of the two code paths that reach the vulnerable query (Bishop Fox, 2026-05-06).

The Sysdig Threat Research Team (TRT) observed the first exploitation attempt 36 hours and seven minutes after the advisory was published to the global database.

The novelty of this finding is the speed and precision of the schema-enumeration attempt, not a confirmed compromise.

Sysdig 2026-04-27
Correctionrun 2026-10-11T1307Z-audittitleheadlinesummaryevent_datetagscvesaffected_productstechniquessourcesevidencesourcing_noteclassificationactionsbody

Sysdig, not Bishop Fox or LiteLLM, reported the attackers' queries against LiteLLM_VerificationToken, litellm_credentials and litellm_config, and the first attempt came 36 hours and seven minutes after the advisory (Sysdig, 2026-04-27); no source gives 26 hours. The flaw is a SQL injection in the key lookup (LiteLLM, 2026-04-20), and Bishop Fox's analysis is dated 2026-05-06 (Bishop Fox, 2026-05-06).

Improvementrun 2026-10-21T1149Z-audittitlesummarybodyevidenceheadline

Sysdig's observation was an exploitation attempt without follow-through: it saw no authenticated calls with exfiltrated keys, no virtual-key minting through /key/generate and no reuse of provider credentials, and calls its finding "the speed and precision of the schema-enumeration attempt, not a confirmed compromise" (Sysdig, 2026-04-27). Sysdig still advises treating any internet-facing instance on a vulnerable version as compromised and rotating every virtual key, master key and provider credential it stores (Sysdig, 2026-04-27).

vulnerability09 May 05:00Zmulti-sourceOpen finding →
NOTABLECVE-2026-31431exploitedupdatedNATOB2

CVE-2026-31431 Copy Fail: Linux kernel AF_ALG page-cache write gives any local user root on mainstream distributions, public exploit, CISA KEV-listed (CVSS 7.8)

Xint (Theori) disclosed "Copy Fail" on 2026-04-29: a flaw in the Linux kernel's algif_aead, reachable through AF_ALG, the socket type that exposes the kernel crypto subsystem to unprivileged userspace, lets a local unprivileged user trigger a deterministic, controlled 4-byte write into the page cache by combining AF_ALG with splice() (Xint, 2026-04-29). Corrupting the cached copy of a setuid binary such as su turns that write into root, and the same exploit script worked unchanged on Ubuntu 24.04, Amazon Linux 2023, RHEL 10.1 and SUSE 16 (Xint, 2026-04-29). The bug comes from a 2017 in-place optimization that mainline removed by reverting algif_aead to out-of-place operation (Xint, 2026-04-29). CERT-EU rates it CVSS 7.8 (CERT-EU, 2026-04-29), CISA added it to its Known Exploited Vulnerabilities catalog on 2026-05-01 (CISA KEV catalog, 2026-05-01), and Ubuntu ships fixed kernels 5.15.0-179.189 for 22.04 and 6.8.0-117.117 for 24.04 (Ubuntu, 2026-04-22); Red Hat states that all its fixes are available (Red Hat RHSB-2026-002). Microsoft reports limited in-the-wild privilege escalation involving su that may be associated with either Dirty Frag or Copy Fail (Microsoft Security Blog, 2026-05-08).

It lets an unprivileged local user trigger a deterministic, controlled 4-byte write into the page cache of any readable file on the system.

Xint (Theori) 2026-04-29

While the affected module cannot be blocklisted due to being built into the kernel, the affected functions themselves can be blocked using the following boot arguments:

Red Hat RHSB-2026-002 2026-04-30
Correctionrun 2026-10-11T1307Z-auditkindtitleheadlinesummaryevent_datetagscvesaffected_productstechniquessourcesevidencesourcing_noteclassificationbodyreferences

CISA added CVE-2026-31431 to its catalog on 2026-05-01 (CISA KEV catalog, 2026-05-01). Ubuntu's fixed kernels are 5.15.0-179.189 for 22.04 and 6.8.0-117.117 for 24.04 (Ubuntu, 2026-04-22); the earlier kernel versions had no source. Microsoft reports limited su-based privilege escalation that "may be indicative of techniques associated with either" Dirty Frag or Copy Fail (Microsoft Security Blog, 2026-05-08); it does not describe the two being used as a complementary pair or an initial access through SSH credential stuffing.

Correctionrun 2026-10-21T1149Z-auditsummarybodysourcesevidence

Blacklisting algif_aead does not work on Red Hat Enterprise Linux: Red Hat states that the module "cannot be blocklisted due to being built into the kernel" and gives the boot argument initcall_blacklist=algif_aead_init to block the affected functions, with initcall_blacklist=af_alg_init as an alternative that blocks the AF_ALG interface itself; it warns of possible performance impact on functionality that uses kernel cryptographic functions, and states that all its fixes are available (Red Hat RHSB-2026-002). Xint also gives blocking AF_ALG socket creation with seccomp as an interim measure (Xint, 2026-04-29).

Builds on: WordPress core's REST batch endpoint + a WP_Query SQL injection chain to unauthenticated RCE on…

vulnerability09 May 05:00Zmulti-sourceOpen finding →
ROUTINECVE-2025-68670updatedNATOA2

CVE-2025-68670: xrdp pre-authentication stack overflow (CVSS 9.1), fixed in 0.10.5 in January; canary-protected builds need a second flaw to exploit

The xrdp project published advisory GHSA-rwvg-gp87-gh6f on 2026-01-27: xrdp before 0.10.5 contains an unauthenticated stack-based buffer overflow in the processing of user domain information during the connection sequence, rated critical with CVSS 3.1 9.1, which could let a remote attacker execute arbitrary code (xrdp, 2026-01-27). The impact is lessened where xrdp was built with stack canary protection, because a second vulnerability would then be needed to leak the canary value (xrdp, 2026-01-27). Kaspersky researchers, who reported the flaw on 2025-12-05, published their technical analysis on 2026-05-08: the vulnerable function copies a domain name of up to 512 bytes into a 256-byte stack buffer, the maintainers fixed it in 0.10.5 and backported the patch to 0.9.27 and 0.10.4.1, and Kaspersky's proof of concept stops at the canary check (Kaspersky Securelist, 2026-05-08). No source reports exploitation.

Correctionrun 2026-10-11T1307Z-audittitleheadlinesummaryevent_dateprioritycvestechniquessourcesclassificationactionsbodyverification

The xrdp project disclosed and fixed this flaw on 2026-01-27 and scores it CVSS 9.1; builds with stack canary protection need a second vulnerability to leak the canary before the overflow can be exploited (xrdp, 2026-01-27). Kaspersky's article of 2026-05-08 is the reporting researchers' technical analysis, not the first disclosure (Kaspersky Securelist, 2026-05-08). The table of other CVEs that closed the earlier text belonged to unrelated findings and carried values their own sources contradict.

vulnerability09 May 05:00Zmulti-sourceOpen finding →
ROUTINECVE-2026-40982updatedNATOA2

CVE-2026-40982: Spring Cloud Config Server unauthenticated directory traversal (CVSS 9.1); open-source fixes only for 4.3.x and 5.0.x

Spring published CVE-2026-40982 on 2026-05-06: a request with a specially crafted URL can lead to a directory traversal in the spring-cloud-config-server module, which serves text and binary files to applications (Spring, 2026-05-06). The flaw affects 5.0.0 to 5.0.2, 4.3.0 to 4.3.2, 4.2.0 to 4.2.6, 4.1.0 to 4.1.9 and 3.1.13 and earlier; open-source fixes are 5.0.3 and 4.3.3, while 4.2.7, 4.1.10 and 3.1.14 are available through enterprise support only, and Spring lists no other mitigation (Spring, 2026-05-06). HeroDevs gives the CVSS 3.1 score as 9.1 (AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:N), unauthenticated and low-complexity (HeroDevs, 2026-05-08). HeroDevs first says an attacker can read or write arbitrary files the server process can access, then that the Config Server's resource endpoints are read-only, so the direct impact is information disclosure of secrets, credentials or any other file the server process can read, although the vector scores integrity High (HeroDevs, 2026-05-08). CERT-FR relayed the advisory together with the companion flaws CVE-2026-40981, CVE-2026-41002 and CVE-2026-41004 (CERT-FR, 2026-05-07). No source reports exploitation.

Correctionrun 2026-10-11T1307Z-audittitleheadlinesummaryevent_dateprioritytagsregionscvesaffected_productstechniquessourcesclassificationactionsbody

CVE-2026-40982 scores CVSS 9.1, not 9.8 (HeroDevs, 2026-05-08), and Spring describes a directory traversal, not code execution (Spring, 2026-05-06). Open-source fixes exist only for the 5.0 and 4.3 branches; the 4.2, 4.1 and 3.1 fixes are available through enterprise support only (Spring, 2026-05-06).

Improvementrun 2026-10-21T1149Z-auditbodyheadline

HeroDevs states that the Config Server's resource endpoints are read-only, so the direct impact of the traversal is information disclosure: secrets, credentials or any other file the server process can read, which in typical deployments includes the secrets the Config Server centralizes for a fleet of microservices (HeroDevs, 2026-05-08). The same post also says an attacker can read or write arbitrary files, and the CVSS vector it gives scores integrity High (HeroDevs, 2026-05-08); Spring's advisory describes only a directory traversal (Spring, 2026-05-06).

vulnerability09 May 05:00Zmulti-sourceOpen finding →
Sources: Spring · CERT-FR · HeroDevs

03Research, reports & policy2 items

ROUTINEupdatedNATOA2

ENISA onboards four new CVE Numbering Authorities under its CVE Root, and seven more move over from MITRE Root

ENISA announced on 2026-05-06 that four organisations joined the CVE Program as CVE Numbering Authorities under ENISA Root and that seven CNAs have moved to it from MITRE Root; more than 90 of the program's 510 CNAs are European and can transfer voluntarily (ENISA, 2026-05-06). ENISA, CVE Root for European entities since November 2025, ties the added capacity to an expected rise in vulnerability discovery driven by frontier AI models (ENISA, 2026-05-06).

Correctionrun 2026-10-11T1307Z-auditkindtitleheadlinesummaryevent_dateprioritytagsentitiestechniquessourcesclassificationbody

ENISA's announcement does not mention the Cyber Resilience Act; it describes ENISA as CVE Root for European entities since November 2025 (ENISA, 2026-05-06). No source says the transfer changes the timing or format of advisories.

policy09 May 05:00Zsingle-source · national CERTOpen finding →
Sources: ENISA
ROUTINEupdatedNATOA2

DENIC .de DNSSEC outage: a faulty key rollover made all .de domains unresolvable for validating resolvers for more than three hours on 5 May 2026

During a routine, scheduled DNSSEC key rollover on the evening of 2026-05-05, DENIC distributed signatures for the .de zone that validating resolvers could not verify, so they treated responses for all .de domains as bogus (DENIC, 2026-05-06). DENIC dates the incident from 21:57, began distributing a correct zone at 00:08 and had fully restored normal operation by 01:15 on 2026-05-06, giving no time zone; read as CEST, that is 19:57 to 23:15 UTC (DENIC, 2026-05-06); Cloudflare saw SERVFAILs from about 19:30 UTC and rolled out its own mitigation at 22:17 UTC through an existing override rule, having no native negative trust anchor mechanism (Cloudflare, 2026-05-06); heise gives 21:43 for the first disruption, which read as CEST is 19:43 UTC (heise online, 2026-05-08), so the three sources give three onsets. DENIC's analysis traces the fault to in-house code around its hardware security modules in the third-generation signing system it put into operation in April 2026: instead of one key pair stored on three HSMs, the software generated three different key pairs, one per HSM, and only one of them matched the pre-published DNSKEY (DENIC analysis, 2026-05-08). Because signatures over NSEC3 records were also unverifiable, even second-level domains without DNSSEC failed to resolve, and DENIC's three permanent validation tools detected the anomaly but their alerts were not processed correctly (DENIC analysis, 2026-05-08). DENIC's final report of 2026-06-11 confirms that initial analysis, rules out any compromise of the signing system and any malfunction of Knot or the HSMs, and explains that the test environment, a single HSM at one location, did not fully execute the rollover-agent code, which misbehaves only when several HSMs are connected (DENIC final report, 2026-06-11).

Updaterun 2026-05-10-001sourcesbody

DENIC published its initial analysis on 2026-05-08 (DENIC analysis, 2026-05-08) (heise online, 2026-05-08). A faulty piece of code in its in-house development generated three different key pairs instead of one for the same key tag, 33834, and only one of their public keys was in the DNSKEY record, so only about a third of the signatures could be validated (DENIC analysis, 2026-05-08). Knot itself is standard software in the process; the defect was in DENIC's own code (DENIC analysis, 2026-05-08). DENIC's validation tools detected the faulty signatures as designed, but their alerts were not processed correctly (DENIC analysis, 2026-05-08).

Correctionrun 2026-10-11T1307Z-auditkindtitleheadlinesummaryprioritytagstechniquessourcesclassificationbodysectors

DENIC gives 21:57 and 01:15 without a time zone; read as CEST, the incident ran from 19:57 UTC until normal operation returned at 23:15 UTC (DENIC, 2026-05-06), and Cloudflare dates the onset to about 19:30 UTC and its own mitigation, an override rule rather than a native negative trust anchor, to 22:17 UTC (Cloudflare, 2026-05-06). DENIC's addendum of 11 May 2026 says no key-tag collision occurred; the software generated one key pair per HSM, and only one matched the published DNSKEY (DENIC analysis, 2026-05-08). No source gives 18 million affected domains, singles out government domains, or mentions .ch.

Correctionrun 2026-10-21T1149Z-auditevent_datesourcesbodyheadlinetitle

The outage began on the evening of 2026-05-05 (DENIC, 2026-05-06). DENIC's final report of 2026-06-11 confirms the findings of its initial analysis of 2026-05-08 and lists its short-term measures: additional alerts based on better visibility of errors, including those in the continuously running testing and validation tools; an accelerated procedure to provide a valid zone backup in an emergency; partial validation of the zone before deployment; no further ZSK rollover until further measures are complete, including a more secure software development process and a test environment with higher coverage; and an external security and process analysis (DENIC final report, 2026-06-11).

policy09 May 05:00Zmulti-sourceOpen finding →

04Updates to prior coverage3 items

NOTABLEexploitedupdatedNATOA2

Poland's ABW: attackers breached five municipal water treatment plants in 2025 and in some cases altered equipment parameters, and hacktivists exploited weak passwords on exposed management panels at municipal sites

First published 2026-05-08 · open finding →

Updaterun 2026-05-09-migratedentitiessectorssourcesbody

UPDATE (originally covered 2026-05-08):

ABW's 2024-2025 activity review, published in Polish on 2026-05-06 with an English edition dated 2026-05-25 (ABW, 2026-05-06), names the five affected facilities: Jabłonna Lacka, Szczytno, Małdyty, Tolkmicko and Sierakowo (ABW, 2026-05-25). It names no actor and no access vector for the water-plant intrusions (ABW, 2026-05-25).

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 2026-05-09-migratedregionssectorssourcestagsbody

UPDATE (originally covered 2026-05-08):

Instructure later attributed the intrusion to a vulnerability in its Free-for-Teacher learning system, and ShinyHunters set a pay-or-leak deadline for 2026-05-12 (The Register, 2026-05-12).

HIGHCVE-2026-6973 +4exploitedupdatedNATOA1

CVE-2026-6973: Ivanti EPMM admin-authenticated RCE exploited in limited attacks, fixed with four further EPMM flaws including unauthenticated Sentry certificate issuance (CVE-2026-5787)

First published 2026-05-08 · open finding →

Updaterun 2026-05-09-migratedregionssectorssourcestagsbody

Ivanti EPMM KEV deadline tomorrow (2026-05-10), European Commission, Dutch DPA, Netherlands Council for the Judiciary, and Finnish Valtori confirmed as exploitation targets in prior Ivanti EPMM zero-day waves; 508 EU on-premises instances remain internet-exposed; credential-chaining risk from January 2026 admin-account compromises elevates urgency.

CERT-FR (CERT-FR, 2026-05-07), Germany's BSI (BSI, 2026-05-07) and NCSC-CH (NCSC-CH Cyber Security Hub, 2026-05-08) published advisories on the May 2026 EPMM update. CERT-FR highlights the reuse of administrator passwords stolen in January as the way attackers reach CVE-2026-6973 (CERT-FR, 2026-05-07), and BSI notes that the required admin credentials may have leaked in the January attacks via CVE-2026-1281 and CVE-2026-1340 (BSI, 2026-05-07).

05Action items3 items

Verification & coverage notes1 run

2026-05-09-migrated · unknown · 17 entries published

Items Dropped from Phase 2 Candidates

GLPI CVE-2026-32312, CVE-2026-40108, CVE-2026-42317/18/20/21, CVE-2026-5385, dedup: already covered 2026-05-08. Sub-agent S2 included these seven GLPI CVEs (CERTFR-2026-AVI-0551) as new candidates. Cross-check against state/cves_seen.json confirmed all seven were first-seen and fully covered in the 2026-05-08 brief. Dropped.

cPanel CVE-2026-29201 / CVE-2026-29202 / CVE-2026-29203, § 3 gate not cleared; embargoed details. S1 flagged these three cPanel CVEs reported by watchTowr. Technical details remain under responsible-disclosure embargo (watchTowr post contained no CVSS score, no exploitation confirmation, and no published patch details). None of the § 3 inclusion gates (CISA KEV, vendor ITW confirmation, pre-auth RCE with PoC, ENISA EUVD CVSS-9+/exploited) were met. Dropped.

Apache CloudStack CVE-2026-25077, post-auth, no KEV, no ITW; § 3 gate not cleared. S1 reported CVE-2026-25077 (Apache CloudStack authentication token handling flaw, CVSS 7.2). Authentication required for exploitation (post-auth admin access needed); no KEV entry; no ITW confirmation. § 3 gate not met. Dropped.

IBM Italy / Salt Typhoon state-actor breach, outside 36 h / 72 h recency windows. S4 reported a Corriere della Sera / Il Sole 24 Ore story on alleged Salt Typhoon compromise of IBM Italy infrastructure. Primary source dates: 2026-05-04 (Il Sole 24 Ore) and 2026-05-05 (BleepingComputer). The 72-hour developing window opened 2026-05-06 00:00 UTC; the primary developments predate this. No material new developments published within the window were identified. Dropped.

ChipSoft (Netherlands healthcare IT), primary event outside window; secondary source unverifiable. S4 flagged a potential ChipSoft breach. The primary development (a ChipSoft advisory) was dated 2026-04-29, outside the 72-hour developing window. A May 7 DataBreaches.net reference was attempted via bridge fetch and returned HTTP 403. With the primary event outside the window and no verifiable secondary source, this was dropped.


Single-Source Items (§ 3 National-CERT Carve-Out and Other Exceptions)

CVE-2025-68670 (xrdp), single source (Kaspersky Securelist). Despite a bridge-assisted fetch sweep across NCSC-CH (no post found), CERT-FR, and BSI, no corroborating advisory was found within the recency window. The vendor (xrdp project) has a corresponding GitHub commit and a release at 0.10.5 confirming the patch, which counts as independent confirmation of the patch but not independent vulnerability analysis. Marked [SINGLE-SOURCE] in § 3.

SEPPmail CVE cluster (CVE-2026-44128 et al.); primary advisory is NCSC-CH (national CERT, carve-out applies) plus vendor release notes. No third-party security researcher write-up was found for this cluster. NCSC-CH is a national CERT, qualifying for the national-CERT single-source carve-out per prompt PD-6. Vendor release notes at the SEPPmail downloads portal independently confirm the CVE assignments and patched version. Marked [SINGLE-SOURCE-NATIONAL-CERT carve-out + vendor] in § 3.

Polish ABW water OT named facility list; ABW annual report only. The five named facilities (Jabłonna Lacka, Szczytno, Małdyty, Tolkmicko, Sierakowo) appear only in the ABW Annual Report 2025. SecurityAffairs coverage cites the ABW report as its source; no independent naming was found. The ABW is a national government security agency, and its annual report constitutes an authoritative primary source. The two-source requirement is met at the level of the core story (ABW annual report + SecurityAffairs coverage), but the specific facility names derive from a single document.


URL Integrity Flags

Kaspersky DAEMON Tools URL, Turkish-language path corrected. The original Kaspersky Securelist URL provided by S1 contained a /tr/ path component (https://securelist.com/tr/daemon-tools-supply-chain-attack/...), indicating a Turkish-locale variant. The English canonical URL (https://www.kaspersky.com/blog/daemon-tools-supply-chain-attack/55691/) was verified live and used in all citations. Readers should use the English-path URL for consistency.

Ivanti hub.ivanti.com advisory URL, authentication wall; national CERT advisory substituted. The vendor advisory URL for CVE-2026-5787/CVE-2026-6973 at hub.ivanti.com requires customer portal login and was not fetchable. CERT-FR CERTFR-2026-AVI-0552 and NCSC-CH post 12548 are cited as primary public-access sources throughout. Readers with Ivanti support portal access should cross-reference the vendor advisory for full patch instructions.

SEPPmail patch version discrepancy, resolved. Sub-agent S1 initially cited SEPPmail fixed version as 15.0.2.1 based on the vendor release notes table. S2 and NCSC-CH advisory 12551 cite 15.0.4 / 15.0.4.1. Investigation: 15.0.2.1 is an earlier branch-maintenance release (different security fix scope); 15.0.4 is the current patch release addressing all six CVEs in the cluster. NCSC-CH's recommended version (15.0.4 / 15.0.4.1) is used throughout as the authoritative remediation target.


Coverage Gaps

ENISA EUVD: JavaScript-rendered; returned empty on all fetch attempts during this run. EUVD could not be used as a secondary confirmation source. This is a recurring infrastructure gap for this routine. coverage_gap: enisa-euvd-inaccessible

CCN-CERT-ES (Spain): Geo-blocked (HTTP 451 / 403) on all fetch attempts including bridge. coverage_gap: ccn-cert-es-geoblocked

CISA advisories (ICS-CERT and standard): CISA domains return HTTP 403 on default UA and require the bridge fetcher (tools/fetch_source.py). Bridge was used for CISA KEV status lookups; specific ICS advisories that may be relevant to OT items were not fully enumerated due to bridge throttling (rate limit encountered on third request in window). coverage_gap: cisa-ics-advisories-partial

Unmatched action items (migrated)

  • Verify Captive Portal is disabled: Device > User Identification > Captive Portal Settings > uncheck Enable Captive Portal.
  • If GlobalProtect is not required, disable it. Confirm with show global-protect-gateway summary.
  • Apply/confirm Threat Prevention content update ≥ 8765-9032 and confirm signatures 95817/95818/95820 are in blocking mode.
  • Hunt for rogue admin account name pattern svc-health-check-[6-digit-numeric] in admin account list (show admins).
  • Review running configuration exports for unexpected changes, particularly pre-shared key material.
  • Upgrade to SEPPmail 15.0.4 (patch 15.0.4.1). Contact SEPPmail support if the update channel is unavailable.
  • If patching is delayed: block source IPs outside admin CIDR from paths /gina/diag/ and /gina/api/v1/admin/ at WAF or network perimeter.
  • After patching: confirm /gina/diag/exec returns HTTP 403/404 from an untrusted IP; confirm /gina/api/v1/admin/config/export returns HTTP 401 without a valid session.
  • Rotate LDAP bind credentials, SMTP relay credentials, and S/MIME key store password regardless of whether exploitation is suspected.
  • Review Tomcat access logs (/var/log/seppmail/access_log.*.txt) for historical access to /gina/diag/ or /gina/api/v1/admin/.
  • Patch Ivanti EPMM to the vendor's current patch level (EPMM 11.12.0.4 or 12.1.0.1 per vendor advisory).
  • If patching is not achievable by 2026-05-10: isolate the admin API (TCP/8443) from internet access at the perimeter firewall.
  • Rotate all admin-level credentials on EPMM instances, including those patched for January 2026 CVE-2026-1281/1340 but where passwords were not rotated after that event.
  • Review device management logs for unexpected device enrollment, profile push, or configuration export events since 2026-04-25.
  • EU organisations: confirm GDPR Article 33 notification obligations, if devices enrolled in EPMM belonged to data subjects, the compromise may trigger a personal data breach notification.
  • Identify all LiteLLM Proxy instances in your environment, including self-hosted, cloud-VM, and container deployments.
  • Update to v1.83.7+: pip install --upgrade litellm or pull updated container image.
  • Treat every upstream API key stored in the proxy database as compromised if the instance was internet-accessible during the exposure window (post-2026-04-29 GHSA publication): rotate OpenAI, Anthropic, Azure OpenAI, Cohere, and all other configured provider keys.
  • Review proxy database access logs for time-delayed injection patterns (multiple requests with anomalous Authorization headers, especially those containing SQL metacharacters or sleep directives).
  • Apply available distribution patches: Ubuntu 22.04/24.04, RHEL 8/9, and CentOS Stream are the priority distros with patches available.
  • For unpatched systems: confirm /proc/sys/kernel/unprivileged_userns_clone is set to 0 on Ubuntu/Debian. On RHEL, confirm user.max_user_namespaces=0 via sysctl.
  • If FIPS mode is not enabled and algif_aead is loadable, check lsmod | grep algif_aead and blacklist if not required: echo "blacklist algif_aead" > /etc/modprobe.d/blacklist-algif.conf && update-initramfs -u.
  • Query EDR/software inventory for DAEMON Tools Lite versions 12.5.0.2421 through 12.5.0.2434.
  • On flagged hosts: check for envchk.exe, processes injected into notepad.exe or conhost.exe, and outbound UDP 443 (QUIC) to non-sanctioned destinations.
  • Update to clean version 12.6.0.2445 if DAEMON Tools Lite is authorised in your environment.
  • If your organisation uses Canvas LMS, confirm with Instructure whether you received an institution notification (Instructure stated affected institutions were notified by 2026-05-01).
  • If notified: assess whether enrolled student or staff data was in scope; evaluate GDPR Article 33 notification obligations (72-hour clock runs from the date Instructure provided confirmation of scope to your institution).
  • Review third-party LTI tool provider access grants in Canvas admin console; revoke service accounts for unused integrations.

Migrated from briefs/2026-05-09.md (v2).