Smarty PHP template engine versions 5.0.0 through 5.8.3 expose a security policy bypass allowing file disclosure via the built-in `stream:` resource type. When Smarty Security is enabled and `Security::$streams` is set to `null` or a restricted allowlist, an attacker able to author or influence a template resource reference can use `stream:php://filter/...` to read arbitrary local files accessible to the PHP process - bypassing the very control intended to block such access. A public proof-of-concept was disclosed in GHSA-rjhh-76wf-8xmw; no active exploitation is confirmed in CISA KEV. The issue is fixed in v5.8.4.
Netty's RedisArrayAggregator codec fails to release retained partial aggregate state when the maxElements limit is exceeded, unlike the symmetric maxNestedArrayDepth branch which correctly calls releaseAndClearDepths() before throwing. In applications that keep the channel alive after a codec exception - a common resilience pattern - an unauthenticated peer can exploit this inconsistency to keep attacker-controlled ByteBuf references pinned in Netty's pooled buffer allocator and cause subsequent messages to be consumed into the stale pre-error aggregate. A public proof-of-concept test is included in the GHSA advisory; fixed in Netty 4.1.136.Final and 4.2.16.Final with no KEV listing at time of analysis.
Hardcoded cryptographic credentials embedded in unencrypted ZTE ZXHN F680 router firmware allow a low-privilege authenticated network attacker to extract integrity verification keys and subsequently bypass application integrity checks on the device. Seven specific firmware versions across the V9.0.10 P1, P4, and P6 branches are confirmed affected per the ZTE advisory. No public exploit identified at time of analysis, and the vulnerability is not listed in the CISA KEV catalog.
Denial of Service in the WP Maps WordPress plugin before 4.9.7 allows any authenticated subscriber to exhaust server resources by exploiting a missing capability check in an AJAX action handler. The flaw, discovered and reported by WPScan, permits a low-privileged user to invoke an unrestricted operation that triggers uncontrolled recursion, potentially rendering the affected WordPress site unavailable. A publicly available proof-of-concept exists, though EPSS remains low (0.14%) and CISA has not listed this in KEV, suggesting exploitation is currently opportunistic rather than widespread.
Unauthenticated cloud credential hijacking in the Templately WordPress plugin (all versions before 3.7.1) allows any remote attacker to overwrite the site administrator's stored cloud service connection with an account under their control. This severs the legitimate administrator's access to the cloud template library and redirects the site's template source to attacker-controlled infrastructure, enabling supply-chain-style content injection against the WordPress admin. A publicly available proof-of-concept exists per WPScan, though EPSS remains very low (0.14%, 3rd percentile), and no CISA KEV listing confirms active widespread exploitation at time of analysis.
Broken Object Level Authorization in MStore API WordPress plugin (all versions before 4.21.0) exposes every WooCommerce order and associated customer PII to any authenticated WordPress user, including low-privileged Subscribers. The vendor-orders endpoint performs no caller-scoping check, meaning a Subscriber can enumerate the entire store's order history. Publicly available exploit code exists per WPScan, though EPSS at 0.13% (3rd percentile) indicates exploitation has not become widespread at time of analysis; no CISA KEV listing confirms active exploitation.
Path traversal in go-git's filesystem-backed reference storage allows a malicious Git server to overwrite arbitrary files within the `.git` directory - including sensitive metadata like `.git/config` and `.git/HEAD` - by advertising crafted reference names containing `..` sequences. Any application using go-git with the `storage/filesystem` package that clones from or fetches untrusted remote repositories is affected; versions prior to 5.19.2 (v5 branch) and 6.0.0-alpha.5 (v6 branch) are vulnerable. No public exploit code or CISA KEV listing has been identified at time of analysis, but the attack is straightforward and requires only that an attacker control a Git server the victim application connects to.
Stored Cross-Site Scripting in the Comments WordPress plugin (all versions before 7.6.60) permits unauthenticated attackers to permanently inject malicious JavaScript into the site database by submitting a crafted URL that the plugin fails to escape before rendering inside an HTML attribute. Any authenticated user - including site administrators - who subsequently views the affected content triggers payload execution in their browser, potentially enabling session hijacking and full site takeover. A publicly available exploit is referenced via WPScan, though EPSS at 0.17% (7th percentile) indicates low observed exploitation volume at time of analysis, and the vulnerability is not listed in CISA KEV.
Reflected XSS in the Subscribe2 WordPress plugin (versions before 10.46) allows an unauthenticated attacker to inject arbitrary JavaScript into a public subscription form by luring a visitor to a crafted link. Exploitation requires the victim to interact with the malicious URL - no server-side authentication is needed - enabling session theft, credential harvesting, or DOM-based actions scoped to the victim's browser. A publicly available proof-of-concept exists per WPScan; EPSS at 0.16% (5th percentile) indicates limited automated exploitation despite the low attack complexity.
Payment integrity bypass in the Subscriptions for WooCommerce plugin (versions before 2.0.1) enables financial fraud by allowing substitution of an arbitrary approved PayPal order token during the WooCommerce order-received flow. The plugin captures a client-supplied token and marks any order paid solely based on a COMPLETED capture status, without verifying that the captured amount matches the order total or that the token belongs to the specific order. Unauthenticated attackers - where guest checkout is enabled - can exploit this to obtain goods or subscription access without paying the actual order price. A publicly available proof-of-concept exists; no CISA KEV listing at time of analysis.
SQL injection in SourceCodester Simple Doctors Appointment System 1.0 exposes the application's database to remote manipulation through the unsanitized 'id' parameter in the admin delete-appointment endpoint. The CVSS 4.0 vector (AV:N/AC:L/PR:N/UI:N) indicates low-complexity remote exploitation with no authentication formally required per the reporter's assessment, though the /admin/ path implies a session context that should be independently verified. A public proof-of-concept is available on GitHub, raising practical exploitation risk despite the product's limited real-world footprint; no active exploitation (CISA KEV) has been recorded.
SQL injection in SourceCodester Photo Share Website 1.0 exposes the unauthenticated user registration endpoint (/social/ajax.php?action=signup) to database manipulation via the email parameter. The CVSS 4.0 vector confirms network-exploitable with no authentication, no complexity, and no special attack conditions, while the E:P modifier reflects publicly available exploit code hosted on GitHub. No active exploitation is confirmed in CISA KEV, but the combination of a zero-barrier attack surface and public POC code means any internet-exposed instance should be treated as immediately at risk.
SQL injection in SourceCodester Photo Share Website 1.0 exposes the login endpoint `/social/ajax.php?action=login` to remote unauthenticated exploitation via the `email` parameter, enabling attackers to manipulate backend database queries. The CVSS 4.0 vector confirms zero-barrier access - no credentials, no complexity, no user interaction required against a default deployment. A public proof-of-concept exploit is available on GitHub, meaningfully lowering the threshold for opportunistic attacks, though no active exploitation has been confirmed in CISA KEV.
Stored cross-site scripting in Meow Gallery WordPress plugin (all versions before 5.5.2) allows any WordPress user holding the Author role or higher to persist a JavaScript payload via unsanitized attachment alt text in linked gallery anchor attributes. Any visitor - including site administrators - who subsequently loads a post containing the affected gallery executes the injected script, enabling session hijacking, credential theft, or malicious redirects. A publicly available proof-of-concept exists per WPScan; EPSS at 0.15% reflects low automated exploitation probability, though the threat to sites with multiple or untrusted author accounts is concrete.
Stored cross-site scripting in BNE Testimonials WordPress plugin (all versions before 2.0.8.2) allows contributor-level authenticated users to inject arbitrary JavaScript into an inline script block via an unescaped shortcode attribute. The payload persists in published content and executes in the browser of any visitor who views the affected page, including administrators. A public proof-of-concept is documented by WPScan; however, the vulnerability is not listed in CISA KEV and EPSS sits at the 5th percentile, indicating low automated exploitation probability.
Quadratic-complexity denial of service in Hono's `languageDetector` middleware (versions 4.12.0-4.12.33) allows unauthenticated remote attackers to exhaust CPU resources by sending HTTP requests with crafted language tags containing large numbers of hyphen-separated subtags. The default middleware configuration accepts language values from query parameters, cookies, and the Accept-Language header without authentication, meaning the vulnerable code path is reachable from the network in any default Hono deployment using this middleware. No public exploit code has been identified and this CVE is not in the CISA KEV catalog, but the attack is trivially constructible from the advisory description; the fix was released in version 4.12.34.
Remote denial-of-service in @sveltejs/kit (versions ≤ 2.70.1) is achievable by any unauthenticated attacker who sends a specially crafted Accept header, triggering quadratic backtracking in the framework's content negotiation regex and causing excessive CPU consumption. The CVSS vector (AV:N/AC:L/PR:N/UI:N) confirms zero-barrier remote exploitation against default SvelteKit deployments, though real-world impact is partly mitigated where platform-level header length limits are in place. No public exploit code has been identified at time of analysis, and this vulnerability is not listed in CISA KEV.
Log flooding via unbounded username input in JupyterHub's form-based login handler allows unauthenticated remote attackers to exhaust logging and disk storage resources on multi-user Jupyter notebook servers running versions prior to 5.5.0. The root cause (CWE-400) is that failed login attempts wrote the full attacker-supplied username - with no length cap - directly into application logs across three distinct code paths: the base login handler, the login POST handler, and the PAM authenticator. No public exploit code has been identified at the time of analysis, and this vulnerability is not listed in the CISA KEV catalog.
TLS credential exposure in ZTE ZXHN F680 fiber gateway firmware allows an adjacent-network attacker to extract unencrypted cryptographic key material - classified under CWE-321 (Use of Hard-coded Cryptographic Key) - enabling potential decryption or impersonation of TLS-protected communications between the device and upstream ISP services. Seven discrete firmware versions are confirmed affected per the ENISA EUVD entry and ZTE's own advisory. No public exploit code or CISA KEV listing is confirmed at time of analysis, and the CVSS adjacency requirement and high complexity rating limit realistic exploitation to technically capable actors with local network proximity.
WP Events Manager for WordPress (all versions before 2.2.5) exposes a payment notification endpoint that accepts fraudulent booking-paid confirmations from unauthenticated attackers. The plugin neither validates that an incoming payment callback originates from the configured merchant nor that the notified amount matches the booking total, enabling any remote actor to mark arbitrary bookings - including other users' bookings - as fully paid without money reaching the merchant. A publicly available exploit exists; the vulnerability is not currently listed in CISA KEV, but the AV:N/AC:L/PR:N/UI:N attack profile means exploitation requires zero privileges and trivial effort.
CAPTCHA bypass in the Simple CAPTCHA with Cloudflare Turnstile WordPress plugin before 1.42.0 allows unauthenticated network attackers to defeat anti-abuse protections on Forminator-integrated forms by exploiting a flawed validation cache keying mechanism. After solving a single legitimate Turnstile challenge, an attacker can replay token-less form submissions during the validation cache window because the cache is bound to an attacker-controlled, reusable request value rather than the one-time challenge token. A publicly available proof-of-concept exists (WPScan); this vulnerability is not currently listed in CISA KEV, and no EPSS score was provided in source data.
Cross-user HTML disclosure in Hono's server-side rendering affects all applications using hono/jsx with memo()-wrapped components that read request-scoped context. On a warm server instance, when two requests share comparator-equal props, the memoized render from the first request is served to the second user verbatim - bypassing per-request context isolation for values obtained via createContext()/useContext(), useRequestContext(), or getContext(). This can leak another user's profile data, CSRF tokens embedded in HTML, or role-gated content. No public exploit or CISA KEV listing exists at time of analysis; the vendor released a confirmed fix in v4.12.34.
Uncontrolled memory consumption in pypdf's /ToUnicode CMap parser allows a remote attacker to cause denial of service by supplying a crafted PDF containing font entries with abnormally large bfrange code or string tokens. All pypdf versions prior to 6.15.0 are affected when performing text extraction operations. No public exploit or active exploitation is identified, but the fix is straightforward and the attack requires only the ability to supply a PDF to an application using the library.
Resource exhaustion in pypdf before 6.15.0 enables denial of service via crafted PDF files containing CID font width entries with maliciously large range values. Any application that processes untrusted PDFs using pypdf's text extraction or font parsing routines is affected, as the library iterates over attacker-controlled CID width ranges without bounds enforcement, consuming excessive CPU and memory. No public exploit code has been identified at time of analysis, and this vulnerability is not listed in the CISA KEV catalog.
Insecure Direct Object Reference (IDOR) in the Subscriptions for WooCommerce WordPress plugin before 2.0.1 allows any authenticated customer to read subscription records belonging to other customers by supplying an arbitrary subscription ID. The plugin fails to enforce ownership checks before returning subscription details - subscribed product, status, and date fields - exposing sensitive order metadata across tenant boundaries. A publicly available proof-of-concept exists per WPScan, but with an EPSS of 0.13% (3rd percentile) and no CISA KEV listing, mass exploitation is not currently observed.
Repeated denial-of-service resets in Aviation VHF Link Control X.25 protocol implementations can be triggered remotely over radio frequency by transmitting malformed or out-of-sequence frames. The resulting reset loop degrades communication availability and increases workload for pilots or air traffic controllers, with potential for reduced situational awareness during critical flight operations. No public exploit has been identified at time of analysis, but the radio-frequency attack vector means any actor with appropriate VHF transmitting equipment and protocol knowledge within range could attempt exploitation without authenticating to the target system.
Denial-of-service via spoofed broadcast control frames over radio frequency can simultaneously disconnect multiple aircraft from air traffic control datalink communications, causing controller overload and delayed clearances. The vulnerability is classified ICS-critical and reported through CISA ICS-CERT (ICSA-26-219-01), indicating a fielded aviation infrastructure component is affected. No public exploit code has been identified at time of analysis, though the attack is described as remotely executable over RF with low privilege requirements, making it a meaningful operational risk at any airport deploying the affected system.
Session termination via injected radio frames threatens Controller-Pilot Data Link Communications (CPDLC), a safety-critical aviation channel used globally between aircraft and air traffic control for ATC clearances, weather, and route updates. Remotely transmitted Unnumbered Disconnect (U DISC) frames or malformed Aviation VHF Link Control (AVLC) frames over VHF radio can forcibly collapse active CPDLC sessions, forcing reversion to voice communication and degrading controller throughput in high-traffic airspace. No public exploit code has been identified at time of analysis; however, the ICS-CERT advisory ICSA-26-219-01 confirms the vulnerability and the aviation safety context elevates operational priority well beyond what the numeric CVSS score alone suggests.
Heap out-of-bounds write in RadiAnt DICOM Viewer is triggered when processing JPEG-compressed pixel data embedded in a crafted DICOM file, potentially enabling arbitrary code execution on the victim's workstation. Reported to CISA ICS-CERT as a medical device advisory (ICSMA-26-218-01), the vulnerability requires no privileges and only passive user interaction - consistent with file-preview or file-open workflows common in clinical imaging environments. The provided CVSS 4.0 score of 5.3 rates impact as availability-only (VA:L), which conflicts materially with the described RCE potential; this discrepancy is significant and should be independently verified. No public exploit code or CISA KEV listing has been confirmed at time of analysis.
Unauthenticated log file exposure in the Webbox component of Tobit Laboratories AG TeamDavid allows remote attackers to retrieve sensitive error messages and internal application details without any credentials. All TeamDavid versions through Rollout 524 are affected per CPE cpe:2.3:a:tobit_laboratories_ag:teamdavid:*:*:*:*:*:*:*:*. No public exploit code or active exploitation has been identified; this vulnerability was disclosed as part of a batch of 22 CVEs uncovered by InfoGuard researchers during a systematic security audit of the platform.
Unauthenticated remote attackers can trigger operator-configured scheduled jobs in Grav CMS when the scheduler-webhook plugin is installed and misconfigured, due to a compound-conditional short-circuit (CWE-303) that skips token validation entirely when no webhookToken is set. The primitive is job-triggering rather than attacker-controlled command injection: RCE is only achievable when an administrator has separately configured a scheduled job that shells out, a condition the attacker cannot influence. No public exploit has been identified at time of analysis, and the vendor-released fix is available in Grav v2.0.9.
Symlink-based path traversal in Smarty's Security::_checkDir() allows an attacker who can place a symbolic link inside a trusted directory (such as a template or upload path) to read arbitrary files accessible to the PHP process, bypassing the configured secure_dir sandbox entirely. Affected are Smarty 5.x versions prior to 5.8.2 and all 4.x releases prior to 4.5.7, when Smarty's Security policy is enabled. The root cause is that _checkDir() resolves paths via string normalization only, while the underlying file_get_contents() call faithfully follows symlinks to their targets outside the sandbox. No public exploit code or CISA KEV listing has been identified at time of analysis.
Unverified password change in Sonatype Nexus Repository 3 allows any account holding the nexus:* permission to overwrite the administrator password by invoking an onboarding-phase endpoint that remains accessible after initial setup is complete. The endpoint enforces access using only the presence of a local onboarding artifact rather than a proper lifecycle state check, enabling a privileged-but-non-admin account to hijack the administrator credential without invalidating existing administrator sessions. No public exploit has been identified at time of analysis and Sonatype has released a fix in version 3.95.0.
Kirby CMS's REST API error handler exposes full server filesystem paths to unauthenticated remote users via unsanitized PHP exception messages. Affected are all Kirby installations on versions prior to 4.9.5 (4.x branch) and 5.0.0 through 5.5.1 (5.x branch) where the REST API has not been explicitly disabled. An attacker triggering a PHP-level error receives the raw exception message - including absolute paths such as '/var/www/kirby-site/content/' - enabling server directory enumeration that lowers the bar for chained attacks. No public exploit or CISA KEV listing has been identified at time of analysis.
Partial arbitrary file read in HashiCorp Consul Community Edition and Enterprise 1.18.21 through 2.0.2 allows a network-accessible, highly-privileged attacker to exfiltrate sensitive credential files from the Consul server host. The flaw is scoped to configurations using the Vault Connect CA provider with JWT or AppRole authentication - a non-default but production-relevant deployment pattern - and requires operator:write permission to exploit. No public exploit or active exploitation has been identified at time of analysis; a vendor-released patch is available in Consul 2.0.3 and corresponding Enterprise builds.
Denial-of-service in the TP-Link Archer A6 V4 router's embedded web server allows an authenticated, adjacent-network attacker to crash the httpd process or the device by sending a malformed request to the HTTP-WRITEOEM handler, which fails to validate input before passing it to flash-write logic. All Archer A6 V4 firmware versions prior to V4_1.15.10 Build 260625 Rel.25447 are affected, per EUVD-2026-54656 and TP-Link's own advisory. No public exploit code or active exploitation has been identified at time of analysis.
Denial-of-service in the TP-Link Archer A6 v4 httpd service allows an adjacent, authenticated attacker to crash the web management process or force a device reboot by exploiting a race condition in the asynchronous systool instruction handling path. All firmware builds prior to V4_1.15.10 Build 260625 Rel.25447 are affected; a vendor-released patch has been issued. No public exploit code exists and the vulnerability is not listed in the CISA KEV catalog - real-world risk is constrained by the adjacent network and high-privilege prerequisites.
Out-of-bounds memory read and write in the Zephyr RTOS SF32LB MPI QSPI NOR flash driver allows a low-privileged thread, when explicitly granted flash device access in a CONFIG_USERSPACE build, to disclose arbitrary CPU-addressable memory or corrupt flash and invalidate cache beyond the mapped flash window. The root cause is a signed/unsigned type confusion in the offset bounds check: a crafted negative off_t value converts to a large unsigned integer and the subsequent addition wraps to a small result that passes the check, allowing memcpy to read from or program flash at attacker-controlled addresses. No public exploit code has been identified at time of analysis; exploitation is nonetheless mechanically straightforward for any attacker familiar with signed/unsigned integer confusion vulnerabilities.
Datadog Android app versions prior to 5.9.2 structurally associate authenticated users' Datadog UUIDs with Google Firebase Installation IDs on every successful login, transmitting this stable cross-session identifier - alongside crash stack traces that may include in-flight application data, internal class names, package names, and breadcrumbs - to Google's Firebase Crashlytics backend without an identifiable user consent gate or opt-out mechanism. The exposure is not triggered by an external attacker but occurs automatically during normal authenticated app usage for all affected users. No active exploitation has been identified (SSVC Exploitation: none; no CISA KEV listing), and no public exploit code exists; the primary risk is regulatory and privacy-related rather than direct technical compromise.
Ghostfolio's Stripe checkout callback endpoint unconditionally grants a one-year Premium subscription to any authenticated user who supplies a valid Stripe Checkout Session ID, regardless of whether payment was completed. Versions prior to 3.4.0 retrieve the session object by ID and immediately create a subscription using the session's `client_reference_id` without ever asserting `session.payment_status === 'paid'` or `session.status === 'complete'`. Any registered user of an affected Ghostfolio instance can self-grant Premium access at zero cost, with no public exploit identified at time of analysis but trivial exploitability given the complete description of the missing check.
Authorization bypass in the Stream - Activity Log & Audit Trail WordPress plugin (versions through 4.2.0) exposes the full site audit trail to any authenticated user holding a subscriber-level account or higher. The flaw resides in the plugin's Heartbeat API live-update handler, which returns activity records without verifying the requesting user holds the privileges required to view them (CWE-862). No public exploit or CISA KEV listing exists at time of analysis, but the low exploitation bar - any registered user account suffices - makes this a credible insider or low-effort attacker risk on sites with open registration.
Unvalidated HTTP redirect handling in OPeNDAP Hyrax Docker 1.18.0 enables SSRF-assisted credential exfiltration, leaking Earthdata authentication headers (User-Id, Echo-Token) to attacker-controlled endpoints. The AllowedHosts allowlist - intended to restrict outbound server-side requests - is silently bypassed when Hyrax follows HTTP redirects without re-validating the destination against the allowlist, causing sensitive Earthdata Login (EDL) session headers to be forwarded externally. No active exploitation has been confirmed (EPSS 0.21%, 11th percentile; not in CISA KEV), but the confidentiality impact is rated High due to the direct exposure of bearer-equivalent credentials used to access NASA/Earthdata scientific datasets.
Cross-package resource and model poisoning in NLTK's downloader (versions <= 3.9.4) allows an attacker who can influence what packages a user downloads to silently overwrite trusted corpora and model files with adversarial replacements. The root cause is that nltk.downloader extracts archives into shared namespaces (corpora/, taggers/) rather than package-isolated directories, and performs integrity validation only after extraction is complete - a classic TOCTOU-like design flaw. Poisoned resources survive interpreter restarts and propagate silently into ML pipelines, making this particularly dangerous in reproducibility-sensitive or automated NLP workflows. No public exploit code has been identified at time of analysis and this vulnerability is not listed in CISA KEV.
Forged on-call acknowledgment and confused-deputy execution in the Datadog Android application (versions prior to 5.9.2) allows any co-installed malicious app to silently dismiss real production incident alerts on behalf of the victim by exploiting an unguarded exported Activity with no permission check. An attacker-controlled app can supply crafted Intent extras to forge a backend Acknowledge API call using the victim's authenticated session, redirect the Datadog process into non-exported internal components, and surface a spoofed alert over the device lock screen. No public exploit has been identified at time of analysis; exploitation requires a malicious co-installed app plus a single user tap.
Stored XSS in the Code Embed WordPress plugin (versions prior to 2.6.1) permits a Contributor-role attacker to plant a URL embed token referencing an attacker-controlled server; when an Administrator or Editor previews or reviews the pending post, the plugin fetches the remote URL and injects the unfiltered response body-including arbitrary JavaScript-directly into the rendered page. The root cause is a dual omission in add-embeds.php: no output sanitization on fetched remote content and no unfiltered_html capability check before URL-embed expansion, a pattern confirmed by the commit diff from the fix release. No public exploit has been identified at time of analysis; vendor-released patch 2.6.1 is confirmed by a GitHub release tag, a commit, and the plugin changelog.
Relative path traversal in Dell OpenManage Server Administrator (OMSA) versions prior to 11.1.0.2 enables low-privileged remote attackers to read arbitrary files from the underlying server filesystem. The flaw (CWE-23) affects managed node components across Windows, RHEL 8.10, RHEL 9.4, and SLES 15 deployments, as confirmed by Dell advisory DSA-2026-326. No public exploit code has been identified at time of analysis, and the vulnerability is not listed in the CISA KEV catalog, but the high confidentiality impact in a privileged server management context elevates practical risk beyond the moderate CVSS base score alone.
Path traversal in Phoca Commander (Joomla file manager extension by phoca.cz) versions 1.0.0 through 6.1.3 enables authenticated Joomla administrators to escape intended directory boundaries during delete, copy, and move file operations, reaching arbitrary locations on the host filesystem. The CVSS 4.0 vector confirms high subsequent-system impact (SC:H/SI:H/SA:H) with no direct impact on the extension itself, meaning successful exploitation threatens the broader server environment rather than just the Joomla installation. No public exploit code or active exploitation has been identified at time of analysis.
Notification dismissal in the Datadog Android application (versions prior to 5.9.2) can be triggered by any zero-permission co-installed application through the improperly exported BubbleChatActivity component (CWE-926), allowing silent cancellation of the hard-coded Bits AI chat notification (ID 9201) with no identity validation of the caller or their supplied conversation_id. The impact is limited to nuisance-level denial of that specific notification and potential facilitation of in-app phishing workflows, with chat content remaining server-authentication gated and inaccessible to the attacker. SSVC assessment confirms no known active exploitation, and no public exploit code has been identified at time of analysis.
Stored Cross-Site Scripting in the Ultra Addons for Contact Form 7 WordPress plugin (all versions through 3.5.43) permits authenticated contributors to inject persistent malicious scripts via the Range Slider widget's 'data-label' and 'data-separator' attributes. The Scope:Changed CVSS metric confirms that once injected, the payload executes in the security context of any user who subsequently visits the affected page - including administrators - enabling session hijacking, credential theft, or admin-level action forgery. No public exploit code has been identified at time of analysis, and the vulnerability is not listed in the CISA KEV catalog.
UNC path injection in Tobit TeamDavid's Webbox link-storing endpoint (`//ServerClient_celink.htm`) forces the server to initiate outbound SMB connections to attacker-controlled hosts, leaking NTLM authentication credentials and enabling SMB relay attacks against internal systems. The unvalidated `pathname` parameter accepts arbitrary UNC paths (e.g., `\\attacker\share`), triggering Windows credential-passing behavior that exposes the server's NTLM hash. No public exploit code identified at time of analysis, though the underlying NTLM coercion technique is well-understood and documented as part of a broader audit of 22 CVEs in the David platform published by InfoGuard.