Remote code execution in h2oGPT through 0.2.1 lets unauthenticated remote attackers read, write, and delete arbitrary files by injecting path traversal sequences into the HTTP bearer token of the OpenAI-compatible files API. Because the shipped default API key is empty, authentication is effectively bypassed and the token string is used directly as a filesystem path component, so an attacker can escape the user directory and plant startup hooks or application-loaded files to execute code. Publicly available exploit code exists (reported by VulnCheck); there is no public exploit identified in CISA KEV, and the CVSS 4.0 base score is 9.3 (Critical).
Unauthenticated SQL injection in the Easy Store extension for Joomla (versions 1.0.0 through 2.0.1) by JoomShaper allows remote attackers to inject arbitrary SQL through improperly validated order parameters, yielding full read access to the backend database including credentials and active session tokens. No public exploit has been identified at time of analysis, but the flaw is trivially reachable over the network without authentication and against default installations. Successful exploitation enables complete disclosure of stored user data and hijacking of authenticated sessions.
Arbitrary file read and write in SoftVC VITS Singing Voice Conversion (so-vits-svc, through commit 730930d) lets unauthenticated remote attackers exfiltrate any server-readable file and plant .wav files anywhere the process can write, by supplying attacker-controlled paths in the audio_path field of the /wav2wav Flask route. Because the full-song inference server passes the value verbatim to librosa/torchaudio/soundfile sinks, a single POST can both leak file contents in the HTTP response and drop files on disk. Publicly available exploit code exists (VulnCheck advisory and geo-chen writeup); no public evidence of active exploitation at time of analysis.
Arbitrary file read in PostCSS (npm postcss <= 8.5.11) lets an attacker who controls CSS input embed a `/*# sourceMappingURL=PATH */` comment that the PreviousMap parser dereferences directly against the local filesystem with no scheme, allowlist, or traversal check. Because the read fires under default options (no `from`, `map`, or plugins), any pipeline that processes untrusted CSS is reachable; when the resulting `JSON.parse` SyntaxError is surfaced to the caller it leaks roughly the first 10 bytes of the targeted file, and the same code path yields a file-existence oracle and a DoS primitive. Publicly available exploit code exists (POC), but there is no public evidence of active exploitation, and EPSS is low at 0.50% (40th percentile).
Unauthenticated arbitrary blob creation in Chatwoot before 4.16.0 lets remote attackers abuse the ActiveStorage direct-uploads controller, which lacks authentication (CWE-306), to resolve any tenant's account and conversation and obtain signed PUT URLs for writing attacker-controlled data into the shared storage backend. The flaw affects all self-hosted and multi-tenant Chatwoot deployments prior to 4.16.0, is reported by VulnCheck, and has publicly available exploit code exists (no CISA KEV listing). The CVSS 4.0 vector emphasizes integrity impact (VI:H) with a secondary availability effect (VA:L) and no confidentiality loss.
Unauthenticated arbitrary file read in Microweber CMS through version 2.0.20 lets remote attackers retrieve sensitive files via directory traversal in the static file controller (ServeStaticFileContoller). Because the controller trusts the attacker-controlled ?path= query input over the bound route segment and normalize_path() fails to strip traversal sequences, a single unauthenticated HTTP GET can disclose environment configuration files containing credentials and other system files. Publicly available exploit code exists; the flaw was reported by VulnCheck and carries a CVSS 4.0 base score of 8.7.
Unauthenticated arbitrary file read in Meshery before 1.0.57 lets remote attackers retrieve any file on the host by abusing the /api/system/fileView and /api/system/fileDownload endpoints, which pass a user-controlled file parameter straight to os.Open with no path validation. Reported by VulnCheck and fixed in v1.0.57, it carries a CVSS 4.0 score of 8.7 and requires no authentication or user interaction; no public exploit identified at time of analysis, though the fix commit publicly reveals the exact vulnerable code path.
Arbitrary file write and unauthorized file read in Vanna (the open-source AI SQL agent) through version 2.0.2 allows unauthenticated remote attackers to escape the FileSystemConversationStore base directory by injecting path traversal sequences into the conversation_id parameter of the chat API. Because the affected endpoints require no authentication, an attacker can write attacker-controlled JSON to arbitrary filesystem locations and read conversation metadata from outside the intended store - publicly available exploit code exists and the flaw was reported by VulnCheck. No CISA KEV listing is present, so there is no public exploit identified as being actively used in the wild.
Denial of service in Amazon's aws-smithy-http-server Rust crate (versions prior to 0.66.5) lets remote unauthenticated attackers exhaust server sockets and async tasks by opening many connections and trickling partial requests that never complete. The default serve() path ships without connection read/header timeouts or a concurrent-connection cap, so a classic slowloris-style flood can hang the service indefinitely. No public exploit has been identified at time of analysis, though the technique is well-understood and the CVSS 4.0 base score is 8.7 (High).
Argument injection in the AWS Bedrock AgentCore Python SDK before 1.18.1 lets a remote authenticated user run arbitrary commands inside the Code Interpreter sandbox by embedding shell/argument delimiters in package names passed to install_packages(). Because the CVSS vector is PR:L, an attacker needs some level of authenticated access and the flaw requires user interaction (UI:A) with a high confidentiality and integrity impact (CVSS 4.0 base 8.4). There is no public exploit identified at time of analysis and the issue is not listed in CISA KEV.
Network-segmentation bypass in Pronetiqs (Panduit) IntraVUE 3.2.1a14 and earlier lets a remote, unauthenticated attacker abuse the product as an active proxy/intermediary to relay traffic across otherwise-isolated OT boundaries. Because IntraVUE typically sits with visibility into both IT and industrial control network segments, this confused-deputy behavior effectively neutralizes the network zoning that protects OT assets. Rated CVSS 10.0 with a scope change; no public exploit identified at time of analysis and it is not listed in CISA KEV.
Server-side request forgery in PHPOffice PhpSpreadsheet (>= 4.0.0 through 5.8.0, plus older branches back to 2.0.0) lets an attacker bypass the WEBSERVICE() domain whitelist introduced in 5.4.0 by chaining through an HTTP 302/301 redirect. Because the whitelist only validates the initial URL host while file_get_contents() silently follows redirects without re-validation, an attacker who can upload a crafted XLSX to an app that calls setDomainWhiteList() and getCalculatedValue() can pivot to internal-only targets and read the full HTTP response (up to 32,767 bytes) as a cell value. Publicly available exploit code exists (a full PoC and lab reproduction are in the GitHub advisory); no public exploit identified as actively used in the wild and it is not in CISA KEV.
Denial of service in PhpSpreadsheet (PHPOffice) versions 2.0.0 through 5.8.0 lets remote unauthenticated attackers crash a PHP worker with a ~1 KiB malformed XLS/OLE file. The library's OLERead::read() follows attacker-controlled OLE sector chains without cycle detection, so a sector that points back to itself makes the small-block depot loop append the same data until memory is exhausted. It is triggered during automatic format detection via Reader\Xls::canRead(), so any app that probes untrusted spreadsheet uploads is affected; a working proof of concept is published in the vendor advisory, though the flaw is not in CISA KEV.
Denial-of-service in PhpSpreadsheet's Gnumeric reader lets remote attackers crash PHP worker processes by uploading a tiny gzip-compressed .gnumeric file that decompresses to hundreds of megabytes. The reader's Gnumeric::canRead() calls gzdecode() with no decompressed-size cap during routine file-type detection, so a ~98 KB payload expanding to ~96 MiB exhausts a 64M memory limit and triggers a fatal error before validation. No public exploit is identified at time of analysis, but a working proof-of-concept is described in the advisory and the CVSS score is 7.5 (High) with A:H impact.
Arbitrary .docx read and write in Office-Word-MCP-Server (gongrzhe) through version 1.1.11 lets an attacker who can influence the filename argument passed to the document tools escape the intended working directory using absolute paths or ../ sequences. Because the exposed open/save operations run in an MCP tool context, a malicious or prompt-injected instruction can read sensitive .docx files or create and overwrite documents anywhere the server process has filesystem access. Publicly available exploit code exists (VulnCheck), but there is no public exploit identified as actively used and it is not on CISA KEV.
Denial of service in the brace-expansion Node.js library (through version 5.0.7) lets remote attackers crash a process via memory exhaustion by supplying a short crafted string to expand(). The expand() function capped the result count (default 100,000) but not per-result length, so chaining brace groups keeps the count legal while each string grows unbounded; roughly 7.5 KB of input ('{a,b}'.repeat(1500)) triggers a fatal, uncatchable out-of-memory crash in a default Node.js process. Because the library is reached transitively through minimatch and glob, any app that feeds attacker-influenced strings into brace pattern matching is exposed. There is no CISA KEV entry, but the fix commit and README publish a working proof-of-concept trigger string.
Two-factor authentication bypass in the Security Ninja Premium WordPress plugin (all versions before 5.290) lets a remote attacker who already knows a valid account password log in without supplying the one-time code, defeating enforced 2FA for any user including administrators. One of the plugin's two 2FA code paths never validates the second factor, so the protection can be silently skipped. Publicly available exploit code exists (published by WPScan) and a vendor patch is available; EPSS is low at 0.25% (17th percentile), indicating no evidence of widespread automated exploitation despite the POC.
Broken access control in the Praison AI SEO WordPress plugin before 5.0.7 lets unauthenticated attackers reach several REST API routes that lack authorization checks, allowing them to alter the permalink of any published post and to read the plugin's stored configuration data. WPScan reported the issue and publicly available exploit code exists, though EPSS remains low at 0.19% and it is not on CISA KEV. Exposure of configuration data (potentially including AI-service API keys and SEO settings) plus silent permalink tampering can break site links and SEO ranking with no login required.
Security-policy bypass in the awslabs AWS API MCP Server (versions 0.2.13 through 1.3.46) lets an actor invoke AWS API operations that the operator's configured policy was meant to deny or gate. If the policy-enforcement data fails to initialize at startup, the server continues running with all policy checks silently disabled for the entire life of the process, so deny/gate rules are never applied. There is no public exploit identified at time of analysis and it is not on CISA KEV; the flaw's real-world impact is bounded because the IAM permissions on the configured credentials still apply.
Remote code execution in MZ Automation's libiec61850 IEC 61850 library is possible via a crafted MMS Initiate request that triggers a heap-based buffer overflow (CWE-122); RCE was demonstrated on hosts with ASLR disabled, while ASLR-enabled deployments face memory corruption or denial of service. Any device or application embedding this library and exposing the MMS/TCP (typically port 102) service to attacker-reachable networks is affected. Reported through CISA ICS-CERT (ICSA-26-204-06); no public exploit identified at time of analysis and no EPSS/KEV signal is present in the source data.
Sensitive information disclosure in Pronetiqs IntraVUE (marketed as Panduit IntraVUE) versions 3.2.1a14 and prior allows remote unauthenticated attackers to read the underlying host and network share filesystem through the industrial-network visualization application. Rated CVSS 4.0 9.2 (Critical) by ICS-CERT, the flaw exposes files outside the application's intended access boundary. There is no public exploit identified at time of analysis and it is not listed in CISA KEV, but the network-facing, no-authentication profile makes it a high-priority patch for OT environments.
Out-of-bounds heap read in PyTorch torchvision's native GIF decoder (versions through 0.28.0) lets attackers who supply a malicious or truncated GIF trigger a segmentation-fault denial of service or leak adjacent heap memory. The flaw sits in the libgif read_from_tensor callback, which passes an unclamped length to memcpy. A public exploit and reproducer exist and an upstream fix is available (commit 4e05dc2), but there is no public exploit identified in the sense of confirmed active exploitation - it is not on CISA KEV. Applications that decode untrusted GIFs via torchvision are affected.
Denial-of-service in MZ Automation's lib60870, an open-source IEC 60870-5-101/104 telecontrol protocol library used in ICS/SCADA systems, allows remote attackers to crash the message-parsing process via a malformed protocol frame that triggers an out-of-bounds read (CWE-125). The CVSS 4.0 vector (AV:N/AC:L/PR:N/UI:N) indicates unauthenticated network exploitation with high availability impact and limited confidentiality exposure. There is no public exploit identified at time of analysis and the flaw is not listed in CISA KEV.
Denial of service in MZ Automation's libiec61850 library allows remote unauthenticated attackers to corrupt memory by sending a crafted Read Request, triggering a stack-based buffer overflow (CWE-121). The flaw affects the IEC 61850 MMS server implementation used in power-grid and industrial automation devices, and was disclosed via a CISA ICS advisory (ICSA-26-204-06). No public exploit is identified at time of analysis, and the vendor CVSS 4.0 vector scores only availability impact, indicating a service crash rather than confirmed code execution.
Denial of service in MZ Automation's libiec61850 library allows a network-adjacent attacker to crash an IEC 61850 MMS server by sending a specially crafted WriteRequest whose listOfData field is empty, triggering a NULL pointer dereference (CWE-476) in the Write Named Variable List handler. Any device or application embedding the affected library as an MMS server - commonly protection relays, RTUs, gateways, and SCADA front-ends in electric-power and industrial environments - is exposed. Reported through CISA ICS-CERT (advisory ICSA-26-204-06) with no public exploit identified at time of analysis and no CISA KEV listing.
Cleartext credential exposure in Pronetiqs IntraVUE (Panduit IntraVUE) versions 3.2.1a14 and prior allows an attacker to retrieve stored passwords in plaintext via the product's API, because credentials are never hashed or encrypted at rest. This industrial-network visualization tool is affected across all builds up to and including 3.2.1a14, and per the CVSS 4.0 vector the disclosure is reachable over the network without authentication (VC:H). There is no public exploit identified at time of analysis, and it is not listed in CISA KEV, but the ICS-CERT advisory (ICSA-26-204-04) confirms the flaw.
Deserialization of untrusted data in Johnson Controls victor video management software (versions 2.9 up to but not including 3.0) on Windows allows an adjacent-network attacker to trigger object injection (CAPEC-586) and achieve high impact on confidentiality, integrity, and availability. The CVSS 4.0 vector (AV:A/PR:N/UI:N) indicates unauthenticated exploitation is possible for an attacker positioned on the same local network segment, with no user interaction. There is no public exploit identified at time of analysis and it is not listed in CISA KEV.
Remote code execution in Cal.com's cal.diy product before 5.9.9 allows an unauthenticated attacker to run arbitrary code on the server by sending a crafted React Server Components (RSC) request. The flaw stems from a bundled vulnerable Next.js version (upstream CVE-2025-55182) that unsafely deserializes attacker-controlled input in the React flight protocol. Rated CVSS 4.0 10.0 (critical) with no authentication or user interaction required; no public exploit identified at time of analysis, though the underlying Next.js class of bug is well documented via the GHSA advisory.
Reflected XSS in WP Compress WordPress plugin (versions before 7.10.04) allows unauthenticated remote attackers to execute arbitrary JavaScript in a victim's browser by crafting a malicious link that injects the attacker's domain as the CDN host for script assets. The plugin's loader emits script elements pointing to the attacker-controlled origin without sanitizing the CDN host query parameter, achieving cross-site scope execution within the target WordPress site. Publicly available exploit code exists per WPScan reporting; no active exploitation is confirmed in CISA KEV, and EPSS is low at 0.20% (11th percentile), consistent with SSVC's 'poc, non-automatable, partial impact' classification.
Remote code execution in JetBrains TeamCity before 2026.1.2 and 2025.11.6 allows an attacker to escape the Kotlin DSL sandbox used to define build configurations-as-code and run arbitrary code in the context of the CI server. Because TeamCity servers hold VCS credentials, deployment keys, and build artifacts, a successful escape effectively compromises the entire build pipeline and downstream software supply chain. There is no public exploit identified at time of analysis, and the flaw is not listed in CISA KEV; EPSS is low at 0.40% (33rd percentile).
Unauthorized settings modification in JetBrains IntelliJ IDEA before 2026.2 allows attackers to alter configuration within an active Remote Development session without proper authorization, tagged by the vendor as an authentication bypass. Because Remote Development connects a local client to a remote backend IDE, tampering with settings can subvert the security boundary between client and host. No public exploit identified at time of analysis, and the issue is not listed in CISA KEV, but the vendor-assigned CVSS of 10.0 with scope change signals maximum theoretical severity.
Authentication bypass via unauthorized input injection in JetBrains IntelliJ IDEA before 2026.2 allows remote attackers to inject unauthorized input into an active Remote Development session, undermining the trust boundary between client and backend host. Because the flaw stems from missing authentication (CWE-306) on a Remote Development channel and carries a maximal CVSS 10.0 with scope change, a successful attacker can influence code, files, or execution on the remote workspace. There is no public exploit identified at time of analysis, but the ceiling severity warrants prompt patching.
Path traversal via prompt injection in Void editor through 1.3.4 allows network attackers who can deliver malicious content into the AI agent's processing context to silently read arbitrary host files outside the open workspace, including SSH private keys and cloud credentials. The AI agent's file-reading tools (read_file, ls_dir, get_dir_tree, search_*) accept absolute filesystem paths and file:// URIs without workspace confinement and bypass the built-in approval gate, enabling silent exfiltration. No CISA KEV listing exists, but a publicly available proof-of-concept exploit has been published, confirming exploitability; the AC:H rating reflects the prompt injection prerequisite rather than any technical barrier once malicious content is ingested.
Unauthenticated arbitrary file upload in the rtCamp GoDAM WordPress plugin (versions ≤ 1.12.2) allows remote attackers to plant executable files and likely achieve remote code execution. The flaw lives in the plugin's WPForms video-field integration, which bypasses WordPress's built-in upload sanitization by trusting the client-supplied Content-Type header and moving raw uploads into a web-served directory. No public exploit code has been identified at time of analysis, but the CVSS 9.8 rating and unauthenticated network vector make this a critical-priority patch for any exposed site.
Privilege escalation in Azure Red Hat OpenShift (ARO) lets an already-authenticated network attacker with limited rights (PR:L) bypass authorization controls (CWE-285) to gain elevated privileges with full confidentiality, integrity, and availability impact. Because the CVSS scope is changed (S:C), the compromise can extend beyond the initially authorized boundary into other components of the managed cluster or platform. There is no public exploit identified at time of analysis and it is not on the CISA KEV list, but the 9.9 CVSS and SSVC 'total' technical impact make it a high-priority patch for any ARO tenant.
Remote code execution in Zohocorp ManageEngine ADAudit Plus (all builds before 8606) allows unauthenticated attackers to run arbitrary OS commands by abusing the product's agent API, which fails to sanitize input passed to a shell (CWE-78). Rated CVSS 10.0 by the vendor with a scope-changing vector, the flaw requires no privileges or user interaction and reaches the network-exposed agent communication channel directly. No public exploit identified at time of analysis and it is not listed in CISA KEV, but the maximal severity and unauthenticated nature make it a high-priority patch for any exposed deployment.
Arbitrary file deletion in the Participants Database WordPress plugin (versions <= 2.7.8.3) lets remote, unauthenticated attackers delete any file the web server can access via a path traversal flaw. Deleting critical files such as wp-config.php can force WordPress into its setup routine and enable full site takeover, which is why this carries a CVSS 10.0 rating. There is no public exploit identified at time of analysis and it is not listed in CISA KEV, but the unauthenticated network-reachable nature makes it a high-priority patch.
Code injection in the emarket-design Customer Support Ticket System & Helpdesk (WP-Ticket) WordPress plugin, versions up to and including 6.0.5, lets unauthenticated visitors invoke arbitrary parameterless PHP functions through the 'path' parameter, enabling information disclosure or denial of site functionality. Exploitation is reachable without login because the required WordPress nonce is publicly emitted via wp_localize_script wherever the [emd_form] shortcode is rendered, neutralizing the nonce as an access control. There is no public exploit identified at time of analysis and it is not listed in CISA KEV, though the Wordfence-assigned CVSS is 9.8; the practical impact is narrower than that score because only functions taking no arguments can be called.
Remote code execution in Microsoft Account arises from a heap-based buffer overflow that an unauthorized, network-based attacker can trigger without authentication or user interaction, per the CVSS:3.1 vector (AV:N/PR:N/UI:N). The CVSS temporal data indicates exploit maturity is Unproven (E:U), so no public exploit identified at time of analysis and the issue is not on CISA KEV; however, an official vendor fix (RL:O) with Confirmed report confidence (RC:C) is already available. Given the 9.8 base score and full confidentiality/integrity/availability impact, this warrants priority remediation despite the current absence of exploit code.
Cross-tenant privilege escalation in Casdoor allows a non-global organization admin in one tenant to delete, create, or modify resources in any other tenant by exploiting an authorization mismatch between parameter-based access control and request-body actions. Versions up to and including 3.115.0 are affected. The vulnerability requires low-privilege authentication but leads to full compromise of all tenant data and operations.
Authentication bypass in the miniOrange SAML Single Sign On - SSO Login plugin for WordPress (all versions through 5.4.4) lets unauthenticated attackers log in as any user, including administrators. The flaw stems from mo_saml_validate_signature() applying a loose boolean check to the tri-state integer returned by PHP's openssl_verify(), so the error value -1 is treated as truthy and accepted as a valid signature. Reported by Wordfence with a CVSS of 9.8; no public exploit has been identified at time of analysis, though the root cause and vulnerable code lines are publicly documented in the plugin source.
Server-side template injection in Tugtainer (self-hosted Docker container auto-update tool) versions before 1.30.2 lets any authenticated user achieve root-level OS command execution by injecting Jinja2 expressions into notification templates. The title_template and body_template fields are rendered by an unsandboxed jinja2.Environment, so template payloads escape into arbitrary code running as root inside the container. Rated CVSS 9.9 with a scope change; no public exploit identified at time of analysis, and it is not listed in CISA KEV.
Remote code execution in the WordPress "Advanced Views" plugin (by wplake) affects all versions up to and including 3.8.11, allowing an authenticated user with only Subscriber-level privileges to inject and execute arbitrary code (CWE-94) on the server. Rated CVSS 9.9 by Patchstack because a low-privilege account - trivially self-registered on many WordPress sites - can achieve full compromise with no user interaction. No public exploit identified at time of analysis, and the flaw is not listed in CISA KEV.
Arbitrary code injection in IntelliJ IDEA before version 2026.2 allows remote attackers to execute arbitrary code by convincing a developer to open a malicious UI Designer form file, leading to full system compromise. No active exploitation or public exploit code has been reported, but a critical-severity vendor patch is available.
Remote code execution in beetl 3.20.2 allows unauthenticated remote attackers to execute arbitrary code via the type.new function and property reflection. Public proof-of-concept exploit code exists, though EPSS exploitation probability is low (0.24%) and no active exploitation has been reported to CISA KEV.
Unsafe file extraction in the Joomla GeoIP extension (versions 1.0.0-6.3.8) enables remote code execution by writing files outside intended directories during update processing. The vulnerability (CWE-22) does not require authentication and is triggered when the extension extracts untrusted archive files. EPSS probability of exploitation is low (0.20%), and no active exploitation or public exploit is confirmed at this time.
Exposure of CDN credentials in administrator request URLs in the Cache Cleaner Pro extension for Joomla versions 1.0.0 through
Session-handling weakness in Progress MOVEit Transfer (all builds before 2025.1.5, and 2026.0.0 before 2026.0.3) leaves sessions valid longer than they should be, so an attacker who obtains a still-active session token can access the managed file transfer application without re-authenticating. Progress rates this CVSS 9.8 with total technical impact, though the CWE-613 root cause makes the unauthenticated network vector worth verifying. No public exploit identified at time of analysis, and EPSS is low at 0.18% (8th percentile).
Information disclosure in Progress MOVEit Transfer (all versions before 2025.1.5, and 2026.0.0 before 2026.0.3) stems from an overly permissive cross-domain (CORS) security policy that trusts untrusted origins, allowing a malicious website to read sensitive responses from an authenticated victim's MOVEit session. The vendor rates it CVSS 9.8, though the underlying weakness (CWE-942) and vendor tags point squarely at cross-origin data leakage rather than direct server compromise. No public exploit identified at time of analysis, and EPSS is low at 0.18% (8th percentile).
Server-Side Request Forgery (SSRF) in Cache Cleaner Pro extension for Joomla 1.0.0-9.7.6 allows unauthenticated remote attackers to access internal or reserved network services via crafted custom query URLs. With a CVSS score of 9.8 (AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H) the vulnerability is automatable and has total technical impact, though no public exploit or active exploitation has been detected. EPSS predicts a 0.18% likelihood of exploitation in the next 30 days.