Unauthenticated remote attackers can exploit sqladmin's `ModelView` list view to probe hidden model column values via an ordering oracle, bypassing the UI-level `column_sortable_list` restriction. The `sortBy` HTTP query parameter is resolved through `getattr(model, ...)` and fed directly into ORM `ORDER BY` clauses without server-side validation, meaning any column - including those deliberately hidden from `column_list` - and related-model columns via dotted paths can be sorted. By observing row-order changes across ascending and descending requests, an attacker incrementally infers relative values of unexposed fields. No public exploit code and no CISA KEV listing exist at time of analysis; vendor-released patch is available in 0.27.1.
Session fixation in tirreno versions prior to 0.10.0 allows an attacker who can inject a known session identifier into a victim's browser to hijack that victim's authenticated session. The framework's authentication flow validates credentials and establishes a session but omits the PHP session_regenerate_id() call, so the same token that existed before login remains valid after it. No public exploit code has been identified; exploitation requires a prerequisite attack capability such as network man-in-the-middle access, a companion cross-site scripting flaw, or subdomain cookie injection.
External control of critical state data in Dell Secure Connect Gateway (SCG) 5.0 exposes sensitive information to remote unauthenticated attackers via network-accessible state manipulation. Affected are SCG 5.0 Appliance versions prior to 5.36.00.16 and SCG 5.0 Application versions prior to 5.36.00.00. Impact is limited to partial confidentiality loss (C:L) with no integrity or availability degradation, per CVSS; no public exploit code has been identified and this vulnerability is not listed in CISA KEV.
Script injection in Dell Secure Connect Gateway (SCG) 5.0 exposes organizations running either the Appliance (prior to 5.36.00.16) or Application (prior to 5.36.00.00) builds to unauthenticated remote exploitation via alternate XSS syntax. An attacker can bypass the product's XSS neutralization controls using non-standard or alternate encoding forms, resulting in script injection with limited confidentiality impact per the CVSS C:L assessment. No public exploit code has been identified at time of analysis, and the vulnerability is not listed in the CISA KEV catalog.
Improper neutralization of escape, meta, or control sequences (CWE-116) in Dell Secure Connect Gateway (SCG) 5.0 enables unauthenticated remote attackers to inject malformed sequences into server responses, facilitating phishing attacks against SCG users. Both the Appliance variant (prior to 5.36.00.16) and the Application variant (prior to 5.36.00.00) are affected. Dell has released patched versions per advisory DSA-2026-382; no active exploitation or public exploit code has been identified at time of analysis.
Improper output encoding in Dell Secure Connect Gateway (SCG) 5.0 - both the Appliance (prior to 5.36.00.16) and Application (prior to 5.36.00.00) variants - allows unauthenticated remote attackers to inject unescaped content into application output, enabling phishing campaigns targeting users of the gateway. The vulnerability is classified as CWE-116 (Improper Encoding or Escaping of Output) and carries a CVSS 5.3 medium score, reflecting network-accessible, zero-authentication exploitation with limited confidentiality impact. Dell has released patches for both product variants under advisory DSA-2026-382; no active exploitation or public proof-of-concept has been identified.
Missing authentication on a critical function in Dell Secure Connect Gateway (SCG) 5.0 allows remote unauthenticated attackers to access protected functionality without credentials, resulting in unauthorized information disclosure. Both deployment forms are affected: the Appliance variant prior to 5.36.00.16 and the Application variant prior to 5.36.00.00. The CVSS 5.3 base score (C:L/I:N/A:N) indicates the immediate impact is limited to read-level unauthorized access; no public exploit code or CISA KEV listing exists at time of analysis.
Stored cross-site scripting in Lumi Education UG's h5p-nodejs-library (all versions prior to 10.0.4) allows an authenticated user with H5P content upload access to embed malicious JavaScript inside an H5P package, which then executes in the browsers of any subsequent users who view that content. The CVSS 4.0 vector (SC:H/SI:H) reflects that the primary damage lands on victim browsers - enabling session hijacking, credential theft, or arbitrary DOM manipulation - rather than on the server itself. No public exploit identified at time of analysis, though the CVSS 4.0 E:P supplemental metric indicates exploit code or proof-of-concept exists.
CSV formula injection in Snipe-IT's location-scoping report export allows an authenticated low-privileged user to plant spreadsheet formulas in free-text asset fields that execute when a superuser downloads and opens the report in Excel, LibreOffice Calc, or Google Sheets. The vulnerable endpoint GET /admin/settings/location-scoping-report.csv uses a bare fputcsv() call without the League\Csv\EscapeFormula sanitization that all other CSV exports in the application apply, creating an inconsistency exploitable for data exfiltration via HYPERLINK or WEBSERVICE formulas, or command execution via legacy DDE on Windows Excel. The flaw exists only in master-branch builds after 8.6.3 - it was never included in a tagged release - and is fixed in version 8.7.0.
CSV formula injection in Snipe-IT through 8.6.3 allows authenticated low-privilege users to plant spreadsheet formulas in the unaccepted assets acceptance report export, which execute in the spreadsheet client of any privileged user who downloads and opens the file. The ReportsController::postAssetAcceptanceReport method builds its CSV output by hand without applying the League\Csv\EscapeFormula sanitizer already wired into the six sibling exports in the same controller, permitting injection of =, +, -, @, tab, or CR-prefixed values into free-text fields such as asset name, company name, category, model, or assignee. When a reports.view-privileged user opens the resulting file in Excel, LibreOffice Calc, or Google Sheets, injected functions such as HYPERLINK or WEBSERVICE can exfiltrate spreadsheet context to an attacker-controlled server; on legacy Windows Excel configurations with DDE enabled, arbitrary command execution on the victim's workstation is also achievable. Fixed in version 8.7.0; no public exploit identified at time of analysis.
Deserialization of untrusted data in tile-ai tilelang versions 0.1.0 through 0.1.14 allows an attacker who can write to the kernel cache directory to achieve arbitrary code execution by planting a crafted `params.pkl` file in the cache path. The `KernelCache._load_kernel_from_disk` function previously used cloudpickle to deserialize cached kernel parameters from disk without validation, a classic CWE-502 pattern that permits executing arbitrary Python code at deserialization time. The upstream fix (commit 11ec2397, PR #3143) replaces cloudpickle with TVM's `tvm.ir.save_json`-based JSON serialization, which cannot execute code, but no official patched release has been published at time of analysis.
Silent deletion failure in Snipe-IT before 8.7.0 causes attachment files to persist on disk despite administrators receiving success responses from both the web UI and REST API controllers. The root cause is an unchecked return value from Storage::delete() in UploadedFilesController::destroy() and Api\UploadedFilesController::destroy(), meaning file records are marked deleted and hidden from listings while the physical files remain on the server filesystem. Anyone with direct filesystem or backup access - including system administrators, backup operators, or post-compromise threat actors - can retrieve files an authorized admin believed had been permanently removed. No public exploit or active exploitation has been identified at time of analysis.
Insecure direct object reference in Snipe-IT 7.0.12 through 8.6.3 allows any authenticated user holding the import permission to enumerate all Import records on the instance and retrieve preview data from files uploaded by other users. The Livewire importer component (App\Livewire\Importer) enforces permission only at component mount time but leaves individual record queries unscoped, so calling the selectFile($id) action with an arbitrary auto-incrementing integer bypasses ownership checks. In Full Multiple Companies Support (FMCS) deployments this crosses tenant boundaries; impact is limited to metadata and CSV preview rows (headers and first data row) rather than the complete uploaded file, but those rows commonly contain personal data, asset serials, and license keys. No public exploit or CISA KEV listing has been identified at time of analysis.
Stored XSS in Snipe-IT before 8.7.0 allows authenticated users holding departments.edit permission to plant malicious JavaScript in department names that executes in every department member's browser when they load the My Assets page. The flaw exists because DepartmentPresenter::formattedNameLink() omits HTML encoding in the fallback branch triggered when the viewing user lacks departments.view permission. Vendor-released patch 8.7.0 is available; no public exploit code has been identified at time of analysis.
Samsung Cloud Assistant prior to version 9.0.5 on Android exposes a broadcast receiver that fails to verify the origin of incoming intents, allowing any local application or process to send crafted messages that disable enhanced data protection settings. The attack requires on-device presence - either a co-resident malicious application or physical device access - but demands no special permissions, as reflected in the CVSS 4.0 PR:N metric. No active exploitation has been identified (not in CISA KEV) and no public proof-of-concept is known at time of analysis.
Improper access control in Samsung's SettingsProvider component on Android allows local attackers to read sensitive device settings without holding any special permissions. Affected devices run Android 15, 16, or 17 with Samsung firmware prior to the SMR Sep-2026 Release 1 patch bundle. No public exploit has been identified at time of analysis, and the CVSS 4.0 score of 5.1 with VC:L reflects the limited (though real) confidentiality exposure from a purely local attack surface.
Improper input validation in Samsung Tips prior to Android 17 allows a local, low-privileged attacker to launch arbitrary Android activities under Samsung Tips' privileges. The attack requires user interaction to trigger the flaw, limiting its severity, but successful exploitation grants the attacker access to functionality and data scoped to the Samsung Tips application. No public exploit code or active exploitation has been identified at time of analysis.
Cyrus IMAP before 3.12.4 permits an authenticated sharee who holds the `maySetKeywords` ACL right on a shared mailbox to overwrite that mailbox's RFC 6154 `specialuse` annotation via a JMAP `Mailbox/set` request, a capability that should be restricted to the mailbox owner. By reassigning the shared mailbox to a system role such as Archive or Snoozed, the sharee can cause the victim's mail client to route messages into the now-shared folder, disclosing content beyond what was originally granted. No public exploit has been identified at time of analysis; vendor-released patches are available across three supported stable branches.
Path traversal in the key provider component of Red Hat Build of Keycloak allows a realm administrator to probe arbitrary filesystem paths by supplying crafted keystore parameters, revealing the existence and readability of server-side files. The vulnerability exists because a prior fix for path probing was incomplete, leaving a residual bypass in the keycloak-services library - the core engine shared by both Red Hat Build of Keycloak and Red Hat Single Sign-On 7. No public exploit code has been identified at time of analysis, and exploitation requires realm administrator credentials, materially limiting opportunistic attack surface.
SQL injection in the Mail Mint WordPress email marketing plugin (versions up to and including 1.31.0) allows authenticated attackers holding administrator-level access to extract sensitive records from the WordPress database via the REST API 'status' parameter. The root cause is a well-known $wpdb->prepare() misuse: the attacker-controlled $contact_filter_query fragment is concatenated into the SQL format string before prepare() is called, so prepare()'s placeholder sanitization never processes the injected content. REST API bodies parsed from php://input further bypass WordPress's automatic wp_magic_quotes() escaping, allowing raw double-quote characters to reach the SQL sink unescaped. No public exploit has been identified and the vulnerability is not listed in CISA KEV.
Second-order SQL injection in the WP Crowdfunding WordPress plugin by Themeum (versions ≤2.2.1) enables authenticated shop managers to exfiltrate arbitrary database contents. The attack is split across two phases: the attacker writes a malicious payload once to the `wpneo_reward` post meta via the WooCommerce REST API, and the unsanitized value is then injected into a live SQL query on every subsequent public page view that renders the campaign rewards sidebar - making the extraction persistent and silent. No public exploit or KEV listing is identified at time of analysis, but the high confidentiality impact (C:H) and persistent execution model elevate practical risk beyond the CVSS 4.9 base score.
Cross-site scripting in OpenText Documentum Webtop versions prior to 16.7.1 allows an authenticated attacker to inject malicious scripts into the application that execute in a victim user's browser upon interaction with attacker-controlled content. The CVSS 4.0 vector (PR:L/UI:A) confirms exploitation requires both a valid account and active victim engagement, constraining the realistic attack surface to insider threats or compromised accounts. No public exploit code or active exploitation has been identified at time of analysis.
Address bar spoofing in Google Chrome's CustomTabs feature on Android (prior to 153.0.8010.36) allows a remote attacker to display a falsified URL by leveraging a co-installed application on the victim's device. The flaw, rooted in CWE-451 (User Interface Misrepresentation of Critical Information), enables phishing attacks where users are deceived into believing they are visiting a trusted origin while interacting with attacker-controlled content. Chromium rates this 'Low' severity; EPSS is 0.16% (6th percentile), no public exploits are identified, and no active exploitation is confirmed.
Address bar spoofing in Google Chrome for Android (all versions before 153.0.8010.36) exploits incorrect certificate authorization logic (CWE-863), allowing a remote attacker who can deliver crafted network traffic - such as via a man-in-the-middle position - to cause the browser to display a false origin in the address bar. The primary risk is phishing: users who trust the spoofed address bar may submit credentials or sensitive data to an attacker-controlled endpoint. Chromium's own team rated this Low severity, consistent with its CVSS 4.8 score and AC:H rating; no public exploit code and no active exploitation have been identified, and the EPSS score of 0.12% (2nd percentile) confirms very low real-world exploitation probability at time of analysis.
Improper authorization in the ProxyHandler component of Samsung Android devices allows a locally-executing, low-privileged attacker to read proxy configuration without proper access controls. Devices running Android 14 through 17 are affected and patched in the SMR September 2026 release cycle. No public exploit code exists and no active exploitation has been identified; the CVSS 4.0 score of 4.8 accurately reflects the constrained local-only attack surface.
Out-of-bounds write in Samsung's libcodec2secevrcdec.so codec library exposes Samsung Mobile Devices running Android 14 through 17 to local memory corruption by low-privileged attackers. The flaw resides in a proprietary secure EVR codec decoder and was addressed in Samsung's SMR Sep-2026 Release 1. No public exploit code or active exploitation has been identified at time of analysis.
Out-of-bounds write in the Samsung-proprietary libsaviextractor.so library allows a local low-privileged attacker to corrupt memory on affected Samsung devices running Android 14 through 17. Exploitation requires code execution on the device with at minimum a standard user account, limiting the attack surface to malicious apps or post-compromise scenarios. Samsung has addressed the issue in SMR Sep-2026 Release 1; no public exploit or active exploitation has been identified at time of analysis.
Samsung Mobile Visual Voicemail for Android prior to version 20.1.00.05 improperly exports application components, allowing a co-installed malicious app on the same device to trigger outgoing calls without holding the required Android CALL_PHONE permission. The CVSS 4.0 score of 4.8 (AV:L/PR:L) accurately reflects the local-only, low-privilege attack surface. Samsung has released a patched build (20.1.00.05) in its September 2026 mobile security bulletin, and no public exploit code has been identified at time of analysis.
Permission control bypass in the Huawei HarmonyOS and EMUI Settings module allows an attacker with physical access to a device to read confidential data from a protected scope beyond the Settings component. The vulnerability affects HarmonyOS versions 4.0.0 through 4.3.3 and EMUI versions 14.0.0 through 16.0.0, as confirmed by Huawei PSIRT via the September 2026 security bulletin. The CVSS scope-change indicator (S:C) signals that successfully exploiting the flaw grants access to confidential resources outside the Settings module itself, though integrity and availability are unaffected. No public exploit code and no CISA KEV listing have been identified at time of analysis.
Stored XSS in the decidim-elections Ruby gem (versions before 0.32.0) allows a process-scoped admin or election editor with question-management rights to persist arbitrary JavaScript in election question body fields, which subsequently executes in the browsers of all public visitors and voters accessing affected election pages. The root cause is the `question_title` helper in `application_helper.rb` returning `question.body` marked `.html_safe` without any sanitization boundary, bypassing Rails' automatic HTML escaping. No public exploit code and no active exploitation (CISA KEV) have been identified; the issue was discovered through a formal security audit commissioned by the Decidim Association.
Uninitialized GPU resource use in Google Chrome on Android (prior to 153.0.8010.36) allows remote attackers to read process memory across the sandbox boundary via a crafted HTML page. The CVSS S:C (scope changed) metric confirms the vulnerability crosses Chrome's GPU sandbox, leaking partial memory contents outside the sandboxed GPU process to a network-positioned attacker. No public exploit or CISA KEV listing exists; EPSS of 0.31% (24th percentile) indicates low observed exploitation probability, consistent with Chromium's 'Medium' severity rating.
Out-of-bounds read in Chrome's ANGLE graphics abstraction layer on Windows (prior to 153.0.8010.36) can expose limited memory content outside the Chrome sandbox to a remote attacker via a crafted HTML page. The vulnerability is platform-specific to Windows and requires user interaction to trigger, carrying a Chromium-assigned severity of Low and a CVSS score of 4.7. No public exploit or active exploitation has been identified; EPSS at 0.16% (6th percentile) confirms very low real-world exploitation probability, making this a routine patch priority rather than an emergency.
Insufficient entropy in the PRNG used by Dell Secure Connect Gateway (SCG) 5.0 allows a local low-privileged attacker to predict pseudo-random values - session tokens, authentication nonces, or cryptographic keys - generated by the system, ultimately enabling elevation of privileges. Both the virtual Appliance form factor (prior to 5.36.00.16) and the Application form factor (prior to 5.36.00.00) are affected. No public exploit code or CISA KEV listing has been identified at time of analysis, but the confidentiality-only CVSS vector (C:H/I:N) conflicts with the description's claimed elevation-of-privileges outcome, which warrants independent verification.
Improper authentication in Samsung's ActivityTaskManagerService allows local privileged attackers on affected Android devices to launch arbitrary activities, bypassing intended access controls. The flaw affects Samsung Mobile Devices running Android 14 through 17 and is addressed in the SMR Sep-2026 Release 1 patch. No public exploit code or active exploitation has been identified at time of analysis, and the CVSS 4.0 score of 4.6 reflects the constrained local-privileged attack surface.
LDAP injection in Dell Secure Connect Gateway 5.0 (both Appliance and Application variants) allows attackers to inject malicious LDAP query elements, leading to script injection against the underlying directory service. The CPE data confirms two distinct deployment forms are affected: the virtual Appliance (fixed at 5.36.00.16) and the Application (fixed at 5.36.00.00), both covered under Dell advisory DSA-2026-382. Critically, the description characterizes the attacker as unauthenticated, but the provided CVSS vector specifies PR:H - a direct contradiction that must be clarified with Dell before this vulnerability is triaged or deprioritized.
Hard-coded credentials embedded in Dell Secure Connect Gateway 5.0 (both Appliance and Application deployment forms) allow a remote attacker who discovers those credentials to authenticate without legitimate access and retrieve sensitive information. Dell advisory DSA-2026-382 confirms patches for both variants: Appliance fixed at 5.36.00.16 and Application at 5.36.00.00. No public exploit code or active exploitation (CISA KEV) has been identified at time of analysis.
Hard-coded cryptographic key exposure in Dell Secure Connect Gateway (SCG) 5.0 enables remote information disclosure by allowing an attacker who extracts the embedded key to decrypt protected communications or data stores. Both the Appliance (versions prior to 5.36.00.16) and Application (versions prior to 5.36.00.00) variants of SCG 5.0 are affected per Dell advisory DSA-2026-382. No public exploit code has been identified at time of analysis, and the vendor has released patches.
Unauthorized remote access to Dell Secure Connect Gateway 5.0 is enabled by a hard-coded cryptographic key embedded in both the Appliance and Application editions. An attacker who extracts this key - through binary analysis of the software - can leverage it to gain unauthorized access to protected resources, bypassing the confidentiality controls the cryptographic layer is intended to enforce. No public exploit or active exploitation (CISA KEV) has been identified at time of analysis, but a vendor patch has been issued under Dell DSA-2026-382.
Incorrect authorization in Cyrus IMAP's JMAP snooze feature allows an authenticated low-privileged user to bypass destination-mailbox ACL checks, inserting mail into another user's inbox or any mailbox whose ID is known - even when the attacker holds no insert permissions on that target mailbox. Affected versions are prior to 3.12.4, 3.10.4, and 3.8.8. No public exploit code has been identified at time of analysis, and the flaw is not listed in CISA KEV.
Remote code execution via subscriber-level authorization bypass in the Checkout Custom Fields Builder for WooCommerce WordPress plugin (all versions ≤1.1.5) allows any authenticated subscriber to install an arbitrary attacker-hosted PHP plugin and achieve full server compromise. The plugin registers AJAX endpoints for plugin installation and activation without enforcing WordPress capability checks, and the required nonce is emitted directly into all admin pages visible to subscribers when WooCommerce is inactive - making the exploit trivially accessible to any registered user. No public exploit code has been identified, but the low authentication barrier and clear exploit path represent a critical risk on multi-user WordPress sites.
SQL injection in PbootCMS versions 3.2.0 through 3.2.5 lets any authenticated backend user inject crafted parameters into the User/mod interface to rewrite arbitrary fields — including password hashes and role assignments — of other accounts, resulting in full account takeover. The flaw requires only a low-privileged login (CVSS PR:L) and no victim interaction, so a single compromised or low-privilege account can be escalated into administrative control of the CMS. Exploitation is easy (AC:L) but real-world urgency is tempered: no CISA KEV listing, an EPSS score of just 0.32% (25th percentile), and a public GitHub repository tied to the CVE identifier that suggests exploit code may be circulating.
Heap out-of-bounds read in CUPS's cupsUTF32ToUTF8() transcoding function exposes partial daemon heap memory to adjacent-network attackers without authentication. The vulnerable code path is reachable via SNMP supply-description parsing, meaning any attacker who can spoof or inject crafted SNMP responses on the same network segment as a CUPS host can trigger the read and potentially leak heap contents from the printing daemon. No active exploitation has been identified, and the adjacent-network attack vector significantly limits the exposure surface.
Missing authorization in Google Chrome's Browser component (prior to 153.0.8010.36) permits remote attackers to bypass system access restrictions via a crafted HTML page. The CWE-862 root cause - an absent authorization check rather than a broken one - means the Browser component fails to enforce expected access control boundaries when processing certain HTML content. Chromium rated this 'Medium' severity, consistent with the CVSS 4.3 score reflecting limited integrity-only impact and a requirement for user interaction. No active exploitation has been confirmed and no public exploit code has been identified at time of analysis.
Improper state validation in Google Chrome's Installer component (prior to 153.0.8010.36) allows remote attackers to bypass system access restrictions by serving a crafted HTML page to a user. The vulnerability carries a Chromium-assigned severity of Medium and an NVD CVSS base score of 4.3, with limited integrity impact and no confidentiality or availability consequences. No public exploit code or CISA KEV listing has been identified at time of analysis, and EPSS places exploitation probability at a low 0.30%.
Cross-origin data disclosure in Google Chrome's WebGL subsystem (all versions prior to 153.0.8010.36) enables remote attackers to read residual GPU memory belonging to other origins by serving a crafted HTML page to an unpatched user. The root cause is CWE-908 (use of uninitialized resource): Chrome's WebGL implementation exposed GPU-side buffers to JavaScript before zeroing them, allowing fragments of prior rendering operations from cross-origin contexts to be read back. No active exploitation is confirmed (not in CISA KEV) and EPSS is low at 0.29%, consistent with the Medium Chromium severity rating and the limited, read-only impact.
Authorization bypass in the Eventin event management plugin for WordPress (all versions ≤ 4.1.22) enables attackers to fraudulently claim paid tickets, mark unprocessed orders as completed, deplete ticket inventory, and trigger confirmation emails - all without making payment. While the formal CVSS vector assigns PR:L (subscriber-level access), the CVE description explicitly confirms a fully unauthenticated exploitation path: the wp_rest nonce is emitted on every frontend page, and the order creation endpoint issues an order_access_token to any caller possessing that nonce, granting anonymous users the credential material to invoke the privileged update_booking_status branch. No public exploit code or CISA KEV listing has been identified at time of analysis; a patch changeset is referenced in the WordPress plugin repository.
Cross-site scripting in Google Chrome's SanitizerAPI (all versions prior to 153.0.8010.36) allows a remote unauthenticated attacker to bypass the browser's web origin policy by serving a crafted HTML page to a victim. The SanitizerAPI is Chrome's browser-native HTML sanitization interface; a flaw in its processing allows certain crafted inputs to violate same-origin restrictions, enabling limited cross-origin data exposure. With CVSS 4.3 (Medium), an EPSS of 0.27%, no KEV listing, and no public exploit code identified, this is a low-urgency but confirmed patch target.
Authorization bypass in WP Recipe Maker for WordPress (all versions ≤10.8.0) allows authenticated contributors to manipulate admin-authored recipe content without authorization. Specifically, any contributor-level user can unpublish any admin recipe by overwriting its post_status to draft or pending, and - when the plugin's recipe_use_author setting is configured as 'parent' - can additionally claim ownership of that recipe by substituting their own user ID into post_author. The vulnerability stems from missing authorization checks in the recipe-save code path. No public exploit has been identified and the CVE is not listed in CISA KEV.
Web origin policy bypass in Google Chrome's Loader component (versions prior to 153.0.8010.36) lets a remote attacker circumvent same-origin protections by serving a crafted HTML page, potentially reading cross-origin resources the attacker should not have access to. The CVSS score of 4.3 and Chromium's own 'Medium' severity rating are consistent, reflecting a confidentiality-only impact with no integrity or availability consequence and no sandbox escape. No public exploit code and no CISA KEV listing exist at time of analysis; EPSS places exploitation probability at 0.26% (18th percentile), indicating low real-world exploitation likelihood.
CSS output encoding failure in Google Chrome prior to 153.0.8010.36 allows an unauthenticated remote attacker to potentially bypass the same-origin policy by luring a user to a crafted HTML page, resulting in limited cross-origin information disclosure. The CVSS score of 4.3 aligns with Chromium's own 'Medium' severity classification, reflecting that the impact is scoped to partial confidentiality loss with no integrity or availability consequence. No active exploitation has been confirmed and EPSS remains low at 0.26%, indicating this is currently a theoretical risk addressed by the September 2026 stable channel update.