01CVE-2026-54420: LiteSpeed cPanel plugin root escalation on CloudLinux/CageFS shared hosting, exploited in the wild (CISA KEV). LiteSpeed user-end cPanel plugin CVE-2026-54420 lets a user with FTP or web-shell access escalate to root on CloudLinux/CageFS shared hosting and is actively exploited; its CVE record dates the in-the-wild exploitation to May 2026, and LiteSpeed was alerted on 2026-05-31 (LiteSpeed, 2026-06-01). Added to CISA KEV on 2026-06-15 (CISA, 2026-06-15). Patch to WHM PlugIn 5.3.2.1 (cPanel plugin 2.4.8). →
02WordPress supply-chain compromise via Awesome Motive's CDN backdoors ~1.2M sites. WordPress supply-chain compromise via Awesome Motive's shared CDN tampered OptinMonster / TrustPulse / PushEngage scripts on ~1.2M sites to auto-create rogue admins and a self-hiding backdoor plugin; "update your plugins" did not protect the exposure window. (Sansec, 2026-06-13) →
03PRC UNC6508 ran year-plus espionage through internet-facing REDCap servers and a Google Workspace BCC rule. PRC actor UNC6508 ran year-plus espionage through internet-facing REDCap research servers and abused a Google Workspace content-compliance rule to silently BCC research/defence email to attacker Gmail; REDCap is widely run at Swiss/EU academic medical centres. (Google GTIG, 2026-06-15) →
Google's Threat Intelligence Group attributes a September 2023 – November 2025 espionage campaign to UNC6508, a PRC-nexus cluster that compromised North American academic, medical and military-health organisations by exploiting externally-facing REDCap (Research Electronic Data Capture) servers, then dropping a bespoke PHP implant tracked as INFINITERED (Google GTIG, 2026-06-15). INFINITERED trojanises REDCap's own upgrade mechanism to survive platform updates, harvests credentials from the REDCap login page, and exposes a cookie-gated backdoor for shell, file, SQL and credential operations (Help Net Security, 2026-06-15). The exfiltration tradecraft is the notable part: after pivoting to a Workspace admin account, the actor created a Google Workspace content-compliance rule named "Patroit" that silently BCC-forwarded any message matching ~150 research/defence keywords to an attacker-controlled Gmail address, abusing a legitimate administrative feature rather than dropping exfiltration malware (T1114.003 Email Forwarding Rule), which evades most DLP that watches for new tooling (SecurityWeek, 2026-06-15). Initial access mapped to T1190; web-shell persistence to T1505.003; admin credential reuse to T1078.
Why it matters to us: REDCap is deployed across Swiss and EU university hospitals, cantonal research bodies and clinical-trial coordinators, and the Workspace BCC-rule technique is tenant-agnostic. Hunt now: Google Workspace admin audit logs for content-compliance/BCC rule creation by non-IT-admin accounts (especially rules with external Gmail recipients), and file-integrity-monitor the REDCap upgrade-staging directory and login handlers, standard web-root scanning misses the upgrade-path implant.
Proofpoint details UNK_DeadDrop, a North-Korea-aligned cluster (related to but distinct from Contagious Interview / Famous Chollima) that sent 250+ recruitment-themed phishing emails to ~100 finance, crypto, education and technology organisations over April–May 2026 (Proofpoint, 2026-06-15); the targeted geographies are a US majority followed by the UK, Australia, France, Germany and the Netherlands, among others (The Hacker News, 2026-06-16). The lure links to attacker-controlled GitHub/GitLab repositories carrying a .vscode/tasks.json with runOn: folderOpen; VS Code shows a workspace-trust prompt, but Cursor IDE executes the task silently with no prompt, dropping the open-source Overlord Go C2 that steals browser credentials and crypto wallets (The Hacker News, 2026-06-16). Mapped to T1566.002, T1195.001, T1059.004 and T1555.003.
Why it matters to us: public-sector and fintech development teams that have adopted Cursor are exposed to silent execution on repository open. Hunt for editor processes (code, cursor) spawning shell/script interpreters outside build directories (Sysmon EID 1 parent-image filter); enforce workspace-trust policy and restrict VSIX installation to an approved-publisher allowlist via enterprise policy.
Sansec Forensics found malicious JavaScript appended to the CDN-served api.min.js files shared by three Awesome Motive WordPress plugins, OptinMonster (1.2M+ installs), TrustPulse and PushEngage, injected on 12 June and served from CDN edges into 13 June (Sansec, 2026-06-13). The vendor confirmed the entry point was exploitation of an UpdraftPlus vulnerability on its own marketing server, which leaked the BunnyNet CDN API key used to tamper the scripts (OptinMonster, 2026-06-14). Because the tampering was at the CDN layer and not in the WordPress.org repository, "update your plugins" gave false assurance for the exposure window. The payload waited for a logged-in administrator, then created a hidden admin account and installed a self-hiding backdoor plugin masquerading as "Content Delivery Helper" or "Database Optimizer", concealed from the plugin list, update checks and API responses, beaconing harvested credentials to a tidio.cc lookalike domain (Patchstack, 2026-06-15). Mapped to T1195.002, T1136.001 (create account) and T1027.005 (indicator removal).
Cardiac-monitoring medtech firm iRhythm filed an SEC Form 8-K Item 1.05 on 2026-06-15 reporting that a threat actor used social engineering against business applications hosted by a third party, exfiltrated PHI, PII and proprietary data, and sent a ransom demand on 9 June; the company made its materiality determination on 10 June (SEC EDGAR, 2026-06-15). iRhythm states clinical and device-monitoring systems were unaffected. [SINGLE-SOURCE]; only the SEC primary is available; no independent corroboration yet.
The LiteSpeed user-end cPanel plugin before 2.4.8, bundled with the LiteSpeed WHM PlugIn and fixed in WHM PlugIn version 5.3.2.1, lets a user with FTP or web-shell access on a CloudLinux/CageFS shared-hosting server escalate to root. LiteSpeed says its WHM plugin itself was not affected, that the flaw "is being actively exploited", and that it was alerted on 2026-05-31 (LiteSpeed, 2026-06-01). The CVE description attributes the flaw to mishandled symlinks (The Hacker News, 2026-06-16) and dates its exploitation in the wild to May 2026 (ENISA EUVD, 2026-06-14); LiteSpeed thanks cPanel, which pushed an uninstall command for the user-end plugin on 2026-05-31, for preventing further exploitation on additional servers (LiteSpeed, 2026-06-01). The Hacker News lists a CVSS score of 8.5 (The Hacker News, 2026-06-16). The CVE description places the vulnerable plugin in WHM PlugIn releases before 5.3.2.0 (The Hacker News, 2026-06-16), while LiteSpeed names WHM PlugIn 5.3.2.1 as its fixed release; both put the user-end plugin fix at 2.4.8 (LiteSpeed, 2026-06-01). CISA added it to the Known Exploited Vulnerabilities catalog on 2026-06-15 (CISA, 2026-06-15).
Triage: a match can be a false positive. LiteSpeed's discriminators are generateEcCert immediately followed by packageUserSize for the same user, which legitimate UI flows do not chain, 7 to 10 concurrent calls per attempt where the legitimate UI makes one at a time, and the same source IP hammering both endpoints (LiteSpeed, 2026-06-01).
This vulnerability is being actively exploited, and poses a risk for all user-end plugin versions prior to 2.4.8.
The main text above described cross-account file access and cited NVD for exploitation and a CVSS 8.5 score. LiteSpeed states that a user with FTP or web-shell access can escalate to root and that the flaw "is being actively exploited", fixed in cPanel plugin 2.4.8 and WHM plugin 5.3.2.1 (LiteSpeed, 2026-06-01). It says only its user-end cPanel plugin is affected, not its WHM plugin, and gives 2026-05-31 as the date it was alerted; the CVE record dates the exploitation itself to May 2026 (ENISA EUVD, 2026-06-14). The 8.5 score is now cited to The Hacker News (The Hacker News, 2026-06-16).
The CVE record for CVE-2026-54420 says the flaw was exploited in the wild in May 2026 (ENISA EUVD, 2026-06-14). LiteSpeed gives 2026-05-31 as the date it was alerted, the same day cPanel pushed an uninstall command for the user-end plugin, and thanks cPanel for preventing further exploitation on additional servers (LiteSpeed, 2026-06-01). LiteSpeed's log check therefore applies to servers that have since been upgraded. The CVE description gives the WHM PlugIn boundary as version 5.3.2.0 (The Hacker News, 2026-06-16), while LiteSpeed's fixed release is 5.3.2.1; both put the user-end plugin fix at 2.4.8 (LiteSpeed, 2026-06-01).
CVE-2026-20262 (CVSS 6.5) is an arbitrary file write in the web UI of Cisco Catalyst SD-WAN Manager: the software does not properly validate user-supplied input during a file upload, so an authenticated attacker with at least write access can create or overwrite any file on the system, which Cisco says can later be used to gain root; it affects all deployment types and has no workaround (Cisco PSIRT). Cisco found it in internal testing and became aware of limited exploitation in June 2026 (Cisco PSIRT); CISA added it to its Known Exploited Vulnerabilities catalog on 2026-06-15 as a path traversal (CWE-22) (CISA KEV catalog, 2026-06-15). BleepingComputer reports that Cisco shared no details of the attacks (BleepingComputer, 2026-06-15). First fixed releases are 20.9.9.2, 20.12.7.2, 20.15.4.5, 20.15.5.3, 20.18.3.1 and 26.1.1.2 (Cisco PSIRT), the same first fixed releases Cisco lists for CVE-2026-20245, the netadmin-to-root command injection it disclosed on 2026-06-04 (Cisco PSIRT, CVE-2026-20245 advisory).
In June 2026, the Cisco PSIRT became aware of limited exploitation of this vulnerability.
Cisco's indicators place the planted web application in the application server's deployments directory and record it as a .war deployment (Cisco PSIRT); the earlier text named Tomcat. No cited source attributes exploitation of CVE-2026-20262 to UAT-8616, and Cisco shared no details of the attacks (BleepingComputer, 2026-06-15).
Cisco's indicators name the exact log entries: in /var/log/nms/vmanage-server.log, a SdraAnyConnectFileUploadHandler upload whose path climbs into the WildFly deployments directory /var/lib/wildfly/standalone/deployments/ with a .war archive, followed, though not always, by a WFLYSRV0010 deployment line in vmanage-appserver.log and POST requests to the new application's JSP page in the service-proxy access log; Cisco also says Managers with ports exposed to the internet are at risk (Cisco PSIRT). CISA lists the flaw as a path traversal (CWE-22) (CISA KEV catalog, 2026-06-15). Its first fixed releases are the same as for CVE-2026-20245, disclosed on 2026-06-04 (Cisco PSIRT, CVE-2026-20245 advisory).
phpBB 3.3.17, released on 2026-06-06, fixes two authentication flaws disclosed by Pentest-Tools.com (Pentest-Tools.com, 2026-06-08). CVE-2026-48611 (CVSS 9.4 per Pentest-Tools) lets a single unauthenticated POST to the user control panel's login-link handler with auth_provider=apache route the login through phpBB's Apache provider, which checks only that the HTTP Basic username matches and that the password is non-empty, never comparing the password with the stored hash, and returns a valid session as any active user on default installations (Pentest-Tools.com, 2026-06-08). The CVE record, assigned by HackerOne as numbering authority, scores it CVSS 3.0 9.8 and describes improper authentication checks in phpBB's OAuth implementation that allow account hijacking even when OAuth is not configured or enabled (ENISA EUVD, 2026-06-12); heise reports the same 9.8 (heise online, 2026-06-15). A known username is enough, and a default board's public member list supplies it (Pentest-Tools.com, 2026-06-08). Sources disagree on whether the session also reaches the Administration Control Panel. Pentest-Tools first wrote that the panel's separate password prompt stops the attacker, then added on 2026-06-15 that an attacker can in fact directly access the panel after impersonating an administrator (Pentest-Tools.com, 2026-06-08). Aikido says a password check in front of the panel cannot be bypassed, which is why it rules out direct remote code execution even through the Extension Catalog (Aikido Security, 2026-06-10). Aikido reported the flaw to phpBB on 2026-06-02 and says it affects every version up to 3.3.16 and 4.0.0-a2 (Aikido Security, 2026-06-10), and heise reports that Aikido and Pentest-Tools' researcher found it independently (heise online, 2026-06-15).
CVE-2026-48612 (CVSS 8.3 per Pentest-Tools; 8.0 under CVSS 3.0 in its CVE record (ENISA EUVD, 2026-06-12), the score heise reports (heise online, 2026-06-15)) affects only boards where an administrator has enabled OAuth login with a Google, Facebook or Bitly provider. Missing OAuth state validation and a missing CSRF check on the account-link endpoint let an attacker bind their own identity-provider account to a logged-in victim who opens a crafted link, which gives the attacker a permanent OAuth login to the victim's account (Pentest-Tools.com, 2026-06-08). Pentest-Tools added proof-of-concept requests for both flaws to its write-up on 2026-07-04 (Pentest-Tools.com, 2026-06-08). Neither CVE is in CISA's KEV catalog as of 2026-09-29 (CISA KEV catalog, 2026-09-29).
Triage: the bypass works by forcing the Apache provider instead of the configured one (Pentest-Tools.com, 2026-06-08), so on a board that uses database authentication, a login request naming the Apache provider is the exploit path, not a misconfigured client.
The main text above gave CVE-2026-48611 only the 9.8 score, said no public exploit code existed and advised disabling OAuth as a stop-gap, although the flaw is reachable even where OAuth is not configured or enabled (ENISA EUVD, 2026-06-12). Pentest-Tools.com scores it 9.4 and shows a single unauthenticated request through phpBB's Apache authentication provider that yields a valid session as any active user, administrators included, on default installations of phpBB 3.3.16 and earlier, fixed in 3.3.17 with no configuration workaround. It published proof-of-concept requests on 2026-07-04 (Pentest-Tools.com, 2026-06-08). Pentest-Tools is one of two independent discoverers, with Aikido (heise online, 2026-06-15). CVE-2026-48612 needs OAuth enabled and a victim who opens a crafted link, and it binds the attacker's OAuth identity to the victim's account rather than hijacking a session (Pentest-Tools.com, 2026-06-08). Disabling OAuth helps only against CVE-2026-48612. No cited source says phpBB is common at European universities or municipalities.
The CVSS 9.8 for CVE-2026-48611 and 8.0 for CVE-2026-48612 that heise reports are the CVE records' own CVSS 3.0 scores, assigned by HackerOne as numbering authority (ENISA EUVD, 2026-06-12; ENISA EUVD, 2026-06-12); Pentest-Tools.com, which found both flaws, scores them 9.4 and 8.3 (Pentest-Tools.com, 2026-06-08). The CVE record for CVE-2026-48611 describes improper authentication checks in phpBB's OAuth implementation that allow account hijacking even when OAuth is not configured or enabled (ENISA EUVD, 2026-06-12). That wording stands; what does not hold is disabling OAuth as a stop-gap, which leaves this flaw open.
Varonis Threat Labs disclosed SearchLeak, a three-stage chain in Microsoft 365 Copilot Enterprise Search that Microsoft patched server-side as CVE-2026-42824 (command-injection / information-disclosure, NVD CVSS 6.5) (Varonis, 2026-06-15; Microsoft MSRC). Stage 1: the q URL parameter is passed to Copilot as an executable instruction rather than a sanitised query (parameter-to-prompt injection). Stage 2: an injected <img> tag fires during a streaming-render race before the output sanitiser runs. Stage 3: the exfiltration request is relayed through Bing's server-side image-search fetch (*.bing.com is allowlisted in Copilot's CSP) bypassing the browser CSP and carrying mailbox content, calendar entries, SharePoint/OneDrive files and emailed MFA/OTP codes to an attacker domain, all from a single click on a genuine microsoft.com link (The Hacker News, 2026-06-15). No customer action is required for patched tenants and no in-the-wild exploitation was observed. Mapped to T1566.002 and T1071.001.
Why it matters to us: M365 Copilot Enterprise is in active Swiss-federal and EU public-sector rollouts. The vulnerability class (prompt injection via URL parameter, streaming-render race, and SSRF-relay CSP bypass) will recur in other AI-augmented enterprise apps; build CASB/DLP detection for Copilot search URLs carrying HTML-encoded payloads in the q parameter and for Copilot sessions fetching to non-Microsoft domains.
Obsidian Security published a privilege-escalation-to-RCE chain in LiteLLM (BerriAI), the widely self-hosted AI gateway that proxies 100+ LLM providers behind one OpenAI-compatible API (Obsidian Security, 2026-06-11; The Hacker News, 2026-06-15). The chain starts with CVE-2026-47101 (authorization bypass): the key-generation endpoint accepts a caller-supplied allowed_routes without checking the caller's role, so an internal_user can mint a key reaching admin routes. CVE-2026-47102 (privilege escalation) follows: /user/update lacks field-level authorization, letting any authenticated user set their own user_role to proxy_admin. CVE-2026-40217 (RCE) ends it: LiteLLM's advisory says the sandbox around custom-code guardrails can be escaped using bytecode-level techniques, giving arbitrary code execution in the proxy process, which runs as root in the default Docker image, and that POST /guardrails/test_custom_code needs a proxy-admin credential in default configurations (LiteLLM advisory, 2026-05-07); Obsidian also ties CVE-2026-40217 to the production routes POST /guardrails and PUT /guardrails/{id}, which run user code with exec(), while the vendor's workaround names only the test endpoint (Obsidian Security, 2026-06-11). VulnCheck scores CVE-2026-47102 at CVSS 8.8 (3.1) (The Hacker News, 2026-06-15), and Obsidian rates the chained impact CVSS 9.9; chained, a default low-privilege account reaches the master key, the salt key decrypting stored secrets, the database URL and every configured provider API key, and can rewrite responses delivered to downstream AI agents ("man-in-the-gateway") (Obsidian Security, 2026-06-11). The chain is fixed in v1.83.14-stable, the first release that Obsidian says contains the complete fix set.
CISA added CVE-2026-40217 to its Known Exploited Vulnerabilities catalog on 2026-11-18 on evidence of active exploitation; the alert names no actor or victim and does not say whether attackers held an administrator credential or used the other two flaws (CISA, 2026-11-18).
CISA added CVE-2026-40217 to its Known Exploited Vulnerabilities catalog on 2026-11-18 on evidence of active exploitation it does not describe (CISA, 2026-11-18). LiteLLM's own advisory describes the flaw as a sandbox escape using bytecode-level techniques at POST /guardrails/test_custom_code that gives code execution in the proxy process, root in the default Docker image, requires a proxy-admin credential in default configurations, rates it Moderate at CVSS 4.0 5.9 and offers blocking the endpoint at the reverse proxy as the workaround; it names 1.83.11 as the fix in its text and 1.83.10 in its patched-versions field (LiteLLM advisory, 2026-05-07). v1.83.14-stable remains the fix for the whole chain.
UPDATE (originally covered 2026-06-13): Novo Nordisk published an incident update on 2026-06-15 clarifying the scope of the theft: clinical-trial data taken was pseudonymised (limited direct re-identification risk for trial subjects) (Novo Nordisk, 2026-06-15), but separately stolen healthcare-professional …
Novo Nordisk published an incident update on 2026-06-15 clarifying the scope of the theft: clinical-trial data taken was pseudonymised (limited direct re-identification risk for trial subjects) (Novo Nordisk, 2026-06-15), but separately stolen healthcare-professional (HCP) data was non-pseudonymised, names, registration numbers and contact details (Security Affairs, 2026-06-15).
The non-pseudonymised HCP records bring the incident within GDPR Article 33 breach-notification obligations and raise targeted-phishing risk against named medical professionals (Security Affairs, 2026-06-15). Healthcare and pharma defenders should expect HCP-impersonation and credential-phishing lures referencing the breach.
Council of Europe breached via the Oracle PeopleSoft zero-day (CVE-2026-35273), ShinyHunters claims 297 GB / ~429,000 files and set a 16 June leak deadline; the first European intergovernmental victim named in the 100+-organisation PeopleSoft campaign (§ 4 update). (SecurityWeek, 2026-06-15)
ShinyHunters listed the Council of Europe (the 46-member human-rights body), claiming 297 GB across ~429,000 files, and set a 16 June leak deadline (SecurityWeek, 2026-06-15); The Register reports the listing as part of ShinyHunters' PeopleSoft campaign (The Register, 2026-06-15). This brings a European intergovernmental institution into the 100+-organisation PeopleSoft campaign previously covered as an education-sector wave.
The claimed dataset spans payroll for 10,000+ current and former staff (2011–2026), 14,000+ CVs, and HR records with names, dates of birth, addresses, bank-account, tax/social-security and medical data. The Council of Europe confirmed it "is currently investigating the matter and assessing the situation" and has not confirmed exfiltration (The Register, 2026-06-15; BleepingComputer, 2026-06-15). The vector, unauthenticated HTTP to the /PSEMHUB/hub servlet (T1190), is unchanged; investigate any externally reachable PeopleSoft Environment Management Hub for compromise and block perimeter access to /PSEMHUB/*. Confidence on the victim claim is MEDIUM pending Council of Europe confirmation (extortion-site claim).
Patch every Oracle PeopleSoft instance to a supported PeopleTools release or disable/remove PSEMHUB now (CVE-2026-35273); a WAF rule alone no longer stops this. UNC6240 (ShinyHunters) bypasses literal-string /PSEMHUB/ WAF blocks with the URL-encoded /%50SEMHUB/ path, so a perimeter block that has not been reconfigured to match on the normalized path is not a control. Investigate any instance that was internet-reachable and unpatched at any point since 27 May 2026 for compromise; on any host where a web shell is found, treat the host as compromised and rotate all credentials accessible from the PeopleSoft tier (its service accounts reach PeopleSoft configuration files and database connection strings), starting with hosts where WebLogic runs as root or SYSTEM.
Patch the LiteSpeed cPanel plugin (CVE-2026-54420) by upgrading to LiteSpeed WHM PlugIn version 5.3.2.1, which bundles cPanel plugin 2.4.8, or uninstall the user-end plugin until then. It is actively exploited on shared CloudLinux/CageFS hosting. Prioritise any public-sector tenant on shared hosting.
Confirm M365 Copilot tenants are on the patched build (CVE-2026-42824) and add CASB/DLP detection for Copilot search URLs carrying HTML-encoded q parameters or fetching to non-Microsoft domains.
Pin LiteLLM to v1.83.14-stable or later on every instance and, until it is in, block POST /guardrails/test_custom_code at the reverse proxy or API gateway; on any instance an untrusted user could reach on an older build, rotate every provider API key, the database credentials, any MCP tokens stored in the proxy and the master and salt keys, and audit every Custom Code Guardrail, the litellm_settings.callbacks entries and every proxy_admin holder, because CISA lists CVE-2026-40217 as exploited and the chain reaches those secrets.
Upgrade every internet-reachable phpBB board to 3.3.17: CVE-2026-48611 gives a session as any user on default installations from one unauthenticated request, proof-of-concept requests are public, and no configuration change closes it.
2026-06-16-38d638e1· Claude Opus 4.8 · 12 entries published
Items dropped:
CVE-2026-20251, Splunk Secure Gateway jsonpickle deserialization RCE (CVSS 8.8): did not clear a § 2 inclusion gate, no in-the-wild exploitation, post-auth/low-privilege, surfaced only via an NCSC-NL advisory. Holding for a future brief if exploitation emerges.
Velvet Ant "Operation Highland" (Sygnia, 2026-06-08): already covered in the 2026-W24 weekly summary (long-running campaigns) with the 2026-06-13 daily deep dive on the related Linux-authentication-stack subversion; no in-window (14–16 June) delta, so excluded per PD-8.
FileFix / KongTuke MotW-bypass transition (Intel 471, 2026-06-03): outside the 36 h recency window; the underlying FileFix research predates June 2026. Not pursued.
Astral (Russia) service disruption; Mackay Sugar (AU) incident; Grafana breach claim; Infinite Campus 137k school-staff breach: lower Swiss/EU public-sector relevance; not pursued this run.
Single-source items: iRhythm Holdings breach (§ 1), SEC Form 8-K Item 1.05 primary only; no independent corroboration yet (national-disclosure carve-out does not apply; flagged inline).
Reduced-confidence items: Council of Europe breach (§ 4), extortion-site (ShinyHunters) claim; the Council confirms an investigation but not exfiltration. Confidence MEDIUM pending victim confirmation; the 16 June leak deadline should resolve it by the next cycle.
Contradictions: phpBB CVE-2026-48611 CVSS, Pentest-Tools.com rated it 9.4; NVD assigns 9.8. Brief reports the NVD value. LiteLLM fix timing, sources gave differing fix dates (25 April vs 2 May 2026); brief cites the fixed version (v1.83.14-stable), not a date, to avoid the discrepancy. LiteSpeed CVE-2026-54420 fixed-version, NVD describes the vulnerable range as "before WHM PlugIn 5.3.2.0", while the LiteSpeed vendor advisory states the fix shipped in WHM PlugIn 5.3.2.1 (bundled with cPanel plugin 2.4.8); the brief uses the vendor's 5.3.2.1 as the safe patch target. CVE-2026-48612 CVSS; NVD has not yet scored it; the 8.0 used here is a third-party (HackerOne) score (Pentest-Tools.com assigned 8.3).
Sub-agents: all four research sub-agents (S1–S4, Claude Sonnet 4.6) returned within the wall-clock cap. S2 and S3 completed their research but their findings-YAML writes did not persist on first return; the main agent recovered both files verbatim from the sub-agent transcripts and the run's URL-liveness ledger before composition (no content was fabricated). Verifier: 4 iterations with model rotation (iterations 1 and 3 Claude Opus 4.8; iterations 2 and 4 Claude Sonnet 4.6); verdict CLEAN at iteration 4 after three remediation rounds (phpBB PoC claim, LiteLLM per-CVE CVSS phrasing, LiteSpeed fixed-version 5.3.2.1, UAT-8616 Talos citation, LiteSpeed KEV citation, DPRK targeted-geography attribution, Novo Nordisk HCP-clause citation).
Coverage gaps: inside-it-ch (bridge 403, no unique in-window content); databreaches-net (bridge returned no output); ncsc-ch-weekly-week24 (HTTP 404; Week 24 report not yet published as of run time); anssi-fr-actu (CERT-FR actualité feed stale, no 2026 bulletins); sophos-xops (no new X-Ops research post in-window); rapid7-research (RSS feed returned empty); cnil-fr (no in-window enforcement notices).
Unmatched action items (migrated)
Audit WordPress sites running OptinMonster / TrustPulse / PushEngage active during 12–13 June UTC, hunt for unexpected admin accounts and for plugins present on disk but hidden from the admin list; pin external CDN scripts to Subresource Integrity hashes. See § 1.
Hunt Google Workspace for rogue content-compliance / BCC rules with external Gmail recipients created by non-IT-admin accounts, and file-integrity-monitor the REDCap upgrade-staging directory and login handlers (UNC6508). See § 1.
Hunt editor-spawned shells: code/cursor processes launching shell or script interpreters outside build directories, and enforce VS Code workspace-trust + VSIX allowlist policy (UNK_DeadDrop). See § 1.