Out-of-bounds kernel memory read and write in Imagination Technologies Graphics DDK exposes systems with PowerVR/Rogue GPUs to full local privilege escalation from any non-privileged user process. Affected versions include Graphics DDK 24.2 RTM2, 25.1 RTM2 through 25.3 RTM, and 26.1 RTM1 - spanning multiple release trains. No public exploit or CISA KEV listing exists at time of analysis; EPSS sits at 0.15% (5th percentile), reflecting low current exploitation momentum despite a CVSS 7.8 local-privilege-escalation profile with High impact across confidentiality, integrity, and availability.
Privilege escalation in Imagination Technologies Graphics DDK allows guest VM kernel software to issue crafted commands to GPU Firmware, triggering writes outside the guest's virtualized GPU memory boundary (CWE-823). Affected are multiple DDK release trains including versions 25.1 RTM2 through 25.3 RTM, 24.2 RTM2, 23.2 RTM2, and 1.18 RTM2. No public exploit code exists and CISA has not added this to KEV; however, SSVC assesses total technical impact given successful exploitation yields full C/I/A compromise of the guest environment.
Server-side request forgery in Headroom LLM proxy (versions before 0.36.1) allows unauthenticated remote attackers to redirect proxy upstream requests to arbitrary internal, loopback, link-local, or RFC 1918 destinations - including cloud instance metadata endpoints such as 169.254.169.254 - by supplying a crafted x-headroom-base-url request header. The proxy returns upstream responses verbatim to the caller, disclosing internal service content; critically, it also forwards the original Authorization header to the attacker-designated host, enabling credential harvesting. Exploitation is trivially achievable against deployments using the vendor-shipped docker-compose.yml, which binds 0.0.0.0 and enforces no HEADROOM_PROXY_TOKEN by default - a configuration weakness the server itself warns about at startup. No public exploit or KEV listing is identified at time of analysis; a vendor patch is available in v0.36.1.
Unauthorized ManagedCluster deletion in Red Hat Multicluster Engine (MCE) allows any tenant with standard ClusterClaim permissions to destroy clusters belonging to other tenants or the hub itself, including the hub's local-cluster. The clusterclaims-controller fails to verify ownership before acting on the spec.namespace field, meaning a low-privileged tenant can target arbitrary ManagedClusters by crafting that field and then deleting the ClusterClaim. No public exploit or CISA KEV listing exists at time of analysis, but the attack requires only standard tenant credentials and no special configuration, making it a realistic insider or compromised-tenant threat.
Access control bypass in Combodo iTop prior to 3.2.3 allows authenticated low-privileged users to read data outside their assigned silos via crafted OQL queries. The iTop silo isolation mechanism - which restricts users to specific data partitions - fails to apply visibility filters to classes joined in OQL but not explicitly listed in the SELECT clause, enabling cross-silo confidential data disclosure. No public exploit code or CISA KEV listing exists at time of analysis; the vendor has released a confirmed fix in version 3.2.3.
Username denylist bypass in Limit Login Attempts Security (Limit Login Attempts Reloaded) before version 3.3.5 allows a WordPress user who has been explicitly blocked by an administrator to successfully authenticate by varying the case of their username or by supplying the account's associated email address instead. This is a logic flaw in the plugin's denylist enforcement, entirely defeating the access control that a site administrator intended to impose on specific accounts. A publicly available proof-of-concept is documented in the WPScan advisory; the low CVSS score of 3.7 with AC:H reflects that successful exploitation requires the attacker to already hold valid credentials for the denylisted account.
Remote memory exhaustion in Bluewave Labs Checkmate prior to version 3.9.1 allows unauthenticated network attackers to crash or destabilize the Node.js backend by submitting concurrent oversized multipart uploads to the public POST /api/v1/auth/register endpoint. The in-memory Multer parser buffers uploaded files before registration validation executes, meaning no file-size, file-count, or MIME-type boundary is enforced prior to business-logic rejection - allowing attacker-controlled allocations to exhaust server RAM. No public exploit has been identified at time of analysis, though the attack requires no credentials, no special tooling, and is easily automatable against any internet-exposed deployment.
Out-of-bounds read and one-byte out-of-bounds write in mod_auth_openidc's state-cookie parser (all versions prior to 2.4.19.4) allow unauthenticated remote attackers to crash Apache worker processes by sending a malformed Cookie header. The root cause is a loop termination condition in src/state.c that checked whether a pointer was NULL — a condition that could never fire since the pointer is only ever incremented — rather than checking for the string terminator, causing state-prefixed cookie tokens lacking '=' to scan and write past the end of the token buffer. No CISA KEV listing or confirmed public exploit exists at time of analysis, but the CVSS AV:N/AC:L/PR:N profile and the fully transparent fix commit make this trivially automatable for denial-of-service once widely known.
Unbounded session accumulation in Spring AI 2.0.0's MCP Streamable HTTP server transport enables remote unauthenticated attackers to exhaust server memory, ultimately causing a Denial of Service that disrupts all legitimate clients. Both WebFlux and WebMvc transport variants are affected because neither enforces a session cap nor mandates client authentication by default. No public exploit code has been identified at time of analysis, but the CVSS vector (AV:N/AC:L/PR:N/UI:N) reflects trivially simple exploitation from any network with access to the MCP endpoint.
Execution After Redirect (EAR) in FuyaWeb Internet and Informatics Services ArchitectPanel Web Admin Panel enables unauthenticated network attackers to bypass authentication controls and access protected administrative content through version 28072026. The flaw exploits a server-side logic error where HTTP redirect responses intended to gate access to the login page continue executing and rendering protected page content, which an attacker can read before following the redirect. No active exploitation has been confirmed (not in CISA KEV) and no public exploit code has been identified; the risk is primarily elevated by the unauthenticated, low-complexity network vector.
Information disclosure in Esri Portal for ArcGIS versions 11.5 through 12.0 allows remote unauthenticated attackers to extract sensitive data reflected in HTTP response bodies via crafted requests. The vulnerability (CWE-200) requires no authentication or user interaction per the CVSS:3.1/AV:N/AC:L/PR:N/UI:N vector, meaning any internet-exposed Portal instance in the affected version range is potentially reachable without credentials. No public exploit or KEV listing exists at time of analysis, and SSVC classifies exploitation as none with partial technical impact, indicating limited observed threat activity despite the high confidentiality impact scored by NVD.
Sensitive information reflection in Esri Portal for ArcGIS versions 11.1 through 12.0 exposes confidential content in HTTP response bodies to remote, unauthenticated attackers under specific, difficult-to-reproduce conditions. The vulnerability stems from improper information exposure (CWE-200), where a crafted request can cause the server to echo sensitive data back in its response. No active exploitation has been confirmed, and EPSS probability is low at 0.33%, though CVSS rates confidentiality impact as high given the potential sensitivity of reflected data.
Unauthenticated access to inline images in Combodo iTop prior to version 3.2.3 is protected by a 24-bit secret generated with PHP's non-cryptographic mt_rand(), yielding only ~16.7 million possible values and making brute-force enumeration feasible for any network-reachable attacker. The commit diff confirms the secret was stored as a 6-character hex string (sprintf('%06x', mt_rand(0, 0xFFFFFF))), and a secondary timing side-channel existed because the comparison used != instead of constant-time hash_equals(). No active exploitation is confirmed in CISA KEV, and no public POC has been identified at time of analysis.
User enumeration in Combodo iTop prior to version 3.2.3 allows remote unauthenticated attackers to identify valid accounts by observing distinct error responses in the forgot-password mechanism. The application returned different messages depending on account state - external authentication, fixed password, missing email - making it possible to distinguish valid usernames from invalid ones without any credentials. No public exploit code has been identified at time of analysis, and the issue is resolved in iTop 3.2.3 per the vendor's GitHub Security Advisory GHSA-888g-gv33-xwwx.
Unauthenticated remote access to uploaded documents is possible in Combodo iTop versions prior to 3.2.3 due to a missing authentication check in the ajax.document.php endpoint. The commit diff confirms that LoginWebPage::DoLoginEx() was absent from critical code paths that serve uploaded files, allowing any network-reachable user who obtains a valid document URL to download sensitive content without credentials. No public exploit code has been identified at time of analysis, but the attack is technically trivial given the low-complexity, unauthenticated CVSS vector.
Denial-of-service in Tie::Hash::Regex before 2.0.0 for Perl allows unauthenticated remote attackers to crash any application that performs hash lookups with externally supplied strings. The FETCH, EXISTS, and DELETE methods compiled caller-provided keys as bare regex patterns without an eval guard, causing the Perl process to die on any syntactically invalid pattern such as an unmatched bracket. No public exploit code has been identified, and EPSS sits at just 0.24% (15th percentile), consistent with a niche library with narrow deployment; however, the impact on any affected application is immediate process termination.
Stored HTML injection in Rocket.Chat's Omnichannel Queue panel allows unauthenticated external attackers to render arbitrary, clickable phishing links in the browsers of support agents viewing the queue. All Rocket.Chat deployments with Livechat/Omnichannel enabled and running versions prior to 8.8.0, 8.7.1, 8.6.2, 8.5.3, 8.4.6, 8.3.8, 8.2.8, 8.1.8, or 7.10.15 are affected. An attacker submitting a crafted visitor name via the unauthenticated POST /api/v1/livechat/visitor REST endpoint can inject arbitrary HTML into the agent-facing side panel UI, enabling targeted social-engineering attacks against customer support staff; no public exploit has been identified at time of analysis, and the EPSS score of 0.18% indicates minimal opportunistic interest.
Sensitive OAuth configuration data is exposed to unauthenticated remote actors in Apache CloudStack's OAuth authentication plugin when the OAuth provider listing endpoint is queried. Affected versions span two release branches: 4.19.0.0-4.20.3.0 and 4.21.0.0-4.22.1.0. Despite a CVSS score of 7.5 reflecting unauthenticated network access and high confidentiality impact, EPSS sits at just 0.17% (7th percentile), and no public exploit code or CISA KEV listing has been identified at time of analysis, suggesting limited active exploitation to date.
Sensitive information exposure in Apache CloudStack's OAuth2 authentication plugin and Google OAuth integration allows unauthenticated remote attackers to access high-confidentiality data across two affected version ranges. The flaw is network-reachable with no authentication or user interaction required (CVSS AV:N/AC:L/PR:N/UI:N), making it a genuine High-severity finding for any CloudStack deployment with OAuth2 or Google OAuth enabled. EPSS sits at 0.17% (7th percentile) and no CISA KEV listing exists, suggesting no public exploitation activity at time of analysis; however, the low attack complexity and zero privilege requirement keep real-world risk elevated for affected configurations.
Stored XSS in the Drag and Drop Multiple File Upload for Contact Form 7 WordPress plugin (versions before 1.3.9.9) allows administrator-level users to inject arbitrary JavaScript via an unsanitized plugin setting used as an HTML tag name in front-end output. Any visitor loading a page that renders the upload field will execute the injected script in their browser context. A publicly available exploit exists; the vulnerability is rated CVSS 3.5 due to the high-privilege requirement, though post-authentication stored XSS can still enable session hijacking or malicious redirects against site visitors.
Missing authorization controls across multiple Userdata-related APIs in Apache CloudStack expose cross-tenant data to any authenticated CloudStack user. Affected APIs - deleteUserData, linkUserDataToTemplate, resetUserDataForVirtualMachine, deployVirtualMachine, updateVirtualMachine, and (from 4.21.0.0) deleteCniConfiguration - fail to validate that the requesting account owns the target userdata resource, enabling one tenant to read, delete, or manipulate another tenant's userdata. No public exploit or CISA KEV listing has been identified at time of analysis; EPSS of 0.13% at the 3rd percentile indicates low automated exploitation activity to date, though the multi-tenant cloud infrastructure context makes successful exploitation high-impact for affected operators.
NTLM credential exposure in Git for Windows before 2.55.0.windows.4 allows a malicious remote Git server to capture Windows authentication material by advertising a crafted bundle URI during clone or fetch operations. When transfer.bundleuri is enabled, non-HTTP(S) URIs including bare UNC paths are treated as local filesystem paths, causing Windows to initiate an outbound SMB connection to an attacker-controlled host and expose NTLM hashes. No public exploit code has been identified at time of analysis and no CISA KEV listing is present, but the technique maps directly onto well-documented NTLM capture and relay attack patterns.
Credential leakage in Bluewave Labs Checkmate versions 3.3.0 through 3.9.1 exposes HTTP Authorization secrets to unauthenticated remote visitors when the global showURL setting is enabled. The unauthenticated GET /api/v1/status-page/:url endpoint returns complete monitor objects from statusPageController.ts, including the secret field consumed by HttpProvider.ts as a bearer or basic-auth credential against monitored upstream services. Although the frontend component BaseStatusPage.tsx intentionally omits this value from the UI, the raw JSON API response delivers it in full, allowing any visitor to harvest credentials and pivot attacks against monitored infrastructure. No public exploit code has been identified at time of analysis, and the issue is fixed in v3.9.2.
Reflected XSS in Combodo iTop prior to version 3.2.3 allows an authenticated low-privilege attacker to inject arbitrary JavaScript via the unsanitized `dashboard_id` parameter in the dashboard revert endpoint `/pages/ajax.render.php`. When a crafted URL is opened by a higher-privileged victim, the payload executes in their browser context, enabling session hijacking or unauthorized actions performed under the victim's identity. No CISA KEV listing or public proof-of-concept exploit has been identified at time of analysis.
DLL hijacking in Remote Utilities Host ≤7.7.3.0 allows any local, unprivileged Windows user to achieve NT AUTHORITY\SYSTEM-level code execution by replacing service DLLs during a race window. The installer incorrectly grants BUILTIN\Everyone FULL CONTROL over all DLL files in the installation directory - including libasset32.dll and seven additional codec and multimedia libraries - enabling overwrite by any local account. No public exploit is identified at time of analysis, but the technique is well-established and the race window reliably materializes during software updates, service crashes, or manual restarts.
Stored Cross-Site Scripting in WPForms Pro for WordPress (all versions through 2.0.0.2) allows unauthenticated remote attackers to inject persistent malicious scripts via form text fields, which execute in the browser of any user who views the stored form entry. The vulnerability exploits a subtle interaction between the plugin's custom kses allowlist expansion - permitting iframes with a data-src attribute - and the WordPress core URI sanitizer's blindspot for that attribute, compounded by the plugin's own admin script promoting data-src to a live src. No active exploitation has been confirmed in CISA KEV and no public proof-of-concept has been identified at time of analysis, though the unauthenticated attack surface and administrative execution context make this a credible priority for WordPress site operators.
Cross-origin RRDP URL confusion in FORT Validator through 1.6.7 allows a delegated Certificate Authority sharing the same Trust Anchor Locator to silently poison the validator's download cache, causing the victim CA's Validated ROA Payloads and signed objects to be purged from FORT's routing origin validation output. The resulting VRP gaps - triggered passively by the victim's own routine publication cycle - can enable BGP route hijacking or loss of network reachability for the affected IP prefixes. No public exploit has been identified at time of analysis, but the CVSS 4.0 score of 7.2 with high integrity impact on subsequent routing infrastructure reflects serious operational risk for networks relying on FORT for RPKI-based origin validation.
Broken object-level authorization in Apache CloudStack exposes project role management APIs to Domain Admins operating outside their assigned domain scope. Authenticated Domain Admins can create, update, delete, and enumerate project roles and role permissions for projects belonging to entirely unrelated domains, because the authorization check validates only that the caller holds the Domain Admin role without confirming the target project falls within their domain or any subdomain hierarchy. Affected versions span 4.15.0.0 through 4.20.3.0 and 4.21.0.0 through 4.22.1.0; no public exploit has been identified at time of analysis and EPSS probability is very low at 0.13%.
Shell-command parser bypass in Omnigent's policy layer allows an authenticated or prompt-injected agent to push code to unauthorized GitHub repositories or escape workspace confinement in all versions prior to 0.3.0. The parser in omnigent/policies/builtins/_shell.py fails to extract the real command when it is wrapped with process utilities (timeout, nice, setsid, stdbuf), combined interpreter flags (bash -lc), command substitutions ($(...) or backticks), or the background operator - returning no recognized tokens, causing both the github.py repository/branch allowlist and working_dir.py workspace confinement policies to abstain and implicitly permit the command. No public exploit code has been identified and no CISA KEV listing exists, but the bypass technique is directly derivable from the published advisory and PR diff.
NULL pointer dereference in TP-Link TL-MR6400 v7's HTTP request parsing crashes the management daemon when an unauthenticated adjacent-network attacker sends a specially crafted request containing a malformed session cookie header. Successful exploitation results in a denial-of-service condition, disabling the web management interface and CGI functionality until service recovery - no confidentiality or integrity compromise is possible. No active exploitation has been confirmed (not in CISA KEV), no public proof-of-concept code has been reported, and a vendor patch is available.
Cross-organization report data exposure in OpenPanel allows any authenticated user to read all analytics reports belonging to another organization's dashboard by supplying a mismatched projectId and dashboardId pair to the report.list tRPC procedure. The enforceAccess middleware validated only project membership, not dashboard ownership, while the underlying getReportsByDashboardId query applied no project-scoping filter - enabling a classic IDOR across tenant boundaries. No public exploit has been identified and no CISA KEV listing exists, but the attack requires only a valid session and knowledge of a foreign dashboardId, making it low-friction in multi-tenant deployments.
Cross-tenant document content corruption in PDFio before 1.6.5 allows low-privileged network attackers to silently overwrite one caller's PDF dictionary string values with another caller's data in multi-threaded or pooled-request environments. The defect is in `pdfioDictSetStringf`, which stored a raw pointer to a stack-local formatted string buffer in the document dictionary rather than heap-allocating a copy, leaving the dictionary referencing reclaimed stack memory after the function returned. No public exploit or CISA KEV listing has been identified at time of analysis; vendor-released patch v1.6.5 is confirmed available.
Reflected XSS in Combodo iTop's dashboard save functionality, present in all releases prior to 3.2.3, allows remote unauthenticated attackers to execute arbitrary JavaScript in the browser of an authenticated iTop user. The vulnerability originates in ajax.render.php, where dashboard_id and reload_url parameters were interpolated directly into inline JavaScript without escaping. With a CVSS Scope Changed classification (S:C), successful exploitation can affect resources beyond the iTop origin. No public exploit code or CISA KEV listing has been identified at time of analysis.
Stack-based out-of-bounds write in the TP-Link TL-MR6400 v7 administrative web interface login handler allows an unauthenticated attacker with adjacent network access to crash the web management service by sending a specially crafted malformed HTTP request. The vulnerability (CWE-787) is confined to the login request parsing path and requires no authentication or user interaction, though exploitation is bounded by the adjacent network attack vector. Impact is limited to a denial-of-service condition - temporary loss of management interface access - with no confirmed confidentiality or integrity impact; no public exploit code has been identified and the vulnerability is not listed in the CISA KEV catalog.
Insecure direct object reference in OpenPanel's tRPC report.get procedure exposes any saved report's full configuration to any authenticated instance user. The enforceAccess middleware in packages/trpc/src/trpc.ts conditionally evaluates project membership only when the input carries a projectId or organizationId key; since report.get accepted only a reportId, it bypassed all authorization checks, and the underlying getReportById query performed no project-level scoping. No public exploit or active exploitation (CISA KEV) has been identified at time of analysis, though the attack pattern is trivially reproducible by any authenticated developer familiar with tRPC.
Path traversal in JuiceFS filestore backend (pkg/object/file.go) enables arbitrary file writes outside the configured storage root during sync operations. The path() helper joined incoming object keys onto the storage root via filepath.Join or filepath.Clean without verifying the resolved result remained within that root, affecting Put, Get, Head, Delete, Chmod, Chown, Symlink, and Readlink. An attacker who can plant object keys containing traversal sequences such as ../../etc/cron.d/backdoor into a source bucket - a shared, public, or compromised bucket - causes a juicefs sync to write attacker-supplied content to arbitrary filesystem paths on the operator's host. No public exploit identified at time of analysis; an upstream fix is available via commit 0bcd70b3d and pull request #7425.
Arbitrary file read in YOOtheme Pro (Joomla extension) versions 2.3.0 through 5.0.40 allows privileged authenticated users to access files outside the intended directory by exploiting glob-based pattern injection in the Filesystem source's path filter. The vulnerability requires high-privilege credentials on the Joomla backend, limiting the attack surface to trusted but potentially malicious insiders or compromised administrator accounts. No public exploit code or CISA KEV listing has been identified at time of analysis, though the CVSS 4.0 score of 7.0 reflects the high confidentiality impact from unrestricted file reads.
Information disclosure in the Brushfire video content streaming platform (online.brushfire.com) exposes internal database path strings within HTTP responses to unauthenticated remote users, allowing enumeration of other users' data. The root cause is CWE-497 (Exposure of Sensitive System Information to an Unauthorized Control Sphere), where backend infrastructure details are insufficiently abstracted from client-facing responses. The vendor applied a server-side fix in February 2026; no active exploitation has been confirmed and no public proof-of-concept has been identified.
Unauthenticated arbitrary directory listing in the Gallery element of the YOOtheme Zoo extension for Joomla exposes server-side directory structures to any remote, unauthenticated attacker on sites running Zoo versions prior to 4.1.66. The path traversal vulnerability (CWE-22) allows crafted requests to enumerate directories outside the intended scope of the Gallery element, potentially revealing sensitive filesystem layout including configuration paths, user data directories, or other installed components. No public exploit code has been identified at time of analysis, and the vulnerability is not listed in the CISA KEV catalog.
STIX document import in misp-stix through version 2026.7.8 allows an unauthenticated attacker who can supply a crafted STIX bundle to inject arbitrary MISP attribute metadata - including distribution level, sharing_group_id, and classification tags - into the importing MISP instance. The import pipeline trusted attacker-controlled STIX fields (tool labels for STIX2, document title for STIX1) to decide whether content originated from MISP itself, and then passed the complete x_misp_attributes dictionary without field filtering directly into attribute creation. No public exploit has been identified at time of analysis; patch commits are available upstream from the MISP project.
Remote unauthenticated attackers can bypass mutual TLS client certificate verification on relay endpoints in Checkmk versions prior to 2.5.0p10 by supplying a fixed placeholder identity value in the request URL. Exploitation results in limited integrity and availability impact against the agent receiver relay channel, with no confidentiality exposure. Only Cloud, Ultimate, and Ultimate MT editions are affected, as relay endpoints are not exposed in other editions; no public exploit has been identified at time of analysis.
Vault token disclosure in the N-able PassPortal browser extension (versions before 3.49.6) stems from missing origin validation on postMessage events, classified under CWE-1385. A malicious web page can send a crafted postMessage to the extension listener, causing it to disclose HashiCorp Vault tokens to an untrusted origin - enabling authentication abuse against any Vault instance the user's session is authorized to access. No public exploit code has been identified at time of analysis, but the network-accessible, unauthenticated attack path (UI:A in CVSS 4.0) and high confidentiality impact make this a meaningful credential theft risk for MSP environments.
Hard-coded credentials embedded in VOCALOID6 by Yamaha Corporation allow an attacker to impersonate a legitimate VOCALOID6 Editor client and authenticate to Yamaha's backend activation and content delivery servers. The CVSS:3.1/AV:N/AC:L/PR:N/UI:N vector indicates this is network-exploitable without authentication or user interaction, as the credentials are distributed within publicly available client software. No public exploit code or active exploitation (CISA KEV) has been identified at time of analysis; the advisory was coordinated through JPCERT/CC (JVNVU90210212).
Unauthenticated file upload in J2Store (Joomla e-commerce extension) versions 1.0.0-3.3.20, 4.0.0-4.0.20, and 4.1.0-4.1.5 permits any remote visitor to POST arbitrary files to the upload endpoint without authentication or CSRF validation. A compounding installer manifest defect omits .htaccess and web.config protection for the upload and invoices directories on fresh installs, making uploaded files directly web-accessible over HTTP. The provided CVSS 4.0 vector includes E:A (Exploit Maturity: Attacked), indicating active exploitation is confirmed at the scoring level, though no corresponding CISA KEV entry was present in the supplied intelligence data.
Server-Side Request Forgery in Apache CloudStack's webhook module lets attackers coerce the management server into issuing arbitrary HTTP requests via webhook delivery, reaching internal-only services and cloud metadata endpoints that are otherwise unreachable from outside. It affects Apache CloudStack 4.20.0.0 through 4.20.3.0 and 4.21.0.0 through 4.22.1.0, carries a CVSS 9.1 (high confidentiality and integrity impact), and has no public exploit identified at time of analysis. EPSS is low at 0.14% (4th percentile) and it is not on CISA KEV, indicating no observed active exploitation yet.
Integer overflow in GNU Emacs PBM/PPM/PGM image loader (all versions before 31.0.91) enables heap memory disclosure when a user opens a crafted image file. Signed integer multiplication of image height, width, and channel count wraps to a negative value for sufficiently large inputs, bypassing the allocated-buffer bounds check and allowing the pixel reader to over-read heap contents, which are then rendered on-screen as color values. No active exploitation has been confirmed (not in CISA KEV), and no public exploit code has been identified at time of analysis; however, the VulnCheck-reported CVSS 4.0 score of 6.9 reflects high confidentiality and availability impact under local conditions with user interaction.
Resource exhaustion in Apache CloudStack's scoped global configuration subsystem-spanning versions 4.7.0 through 4.20.3.0 and 4.21.0.0 through 4.22.1.0-allows authenticated low-privilege users to gradually trigger a denial-of-service condition against the management server. The vulnerability (CWE-772) affects multiple plugins including Quota and Host-HA, which allocate resources during configuration operations but fail to release them after their effective lifetime. No public exploit or active exploitation has been identified; vendor-released fixes are available in 4.20.3.1 and 4.22.1.1.
Symlink-following arbitrary file write in ONNX's save_external_data function (onnx/external_data_helper.py) allows a local attacker with write access to a shared directory to redirect a victim's model serialization writes to any file the victim can modify — including ~/.ssh/authorized_keys, cron entries, and application configuration files. All ONNX releases up to and including 1.20.1 (pip/onnx) are affected; the vendor-released patch is 1.21.0. Publicly available exploit code is confirmed via the GHSA advisory, and CISA has published a CSAF advisory (va-26-233-01) covering this issue, indicating heightened government-sector awareness.
Blind server-side request forgery in the Eventin WordPress plugin prior to version 4.1.21 allows contributor-level authenticated users to direct the WordPress server to issue arbitrary outbound HTTP requests by supplying unvalidated webhook URLs on events they do not own. Two distinct control failures compound the issue: webhook URL input is accepted without sanitization, and the plugin performs no event ownership check, meaning any contributor can abuse another user's event configuration. A publicly available proof-of-concept exists; no CISA KEV listing confirms active exploitation at time of analysis.
Unauthorized post data disclosure in the Duplicate Post WordPress plugin before 1.5.6 allows authenticated users holding a delegated role to read content, metadata, and passwords for posts they should not have access to - including other users' private and draft posts. The CVSS vector (PR:H/C:L) reflects a constrained impact limited to information disclosure within authenticated contexts, making this a meaningful risk in multi-user editorial WordPress environments but not in single-author deployments. A publicly available proof-of-concept exists per WPScan's advisory; no active exploitation has been confirmed via CISA KEV.