CVE Feed

There is a memory corruption vulnerability recently discovered in NI LabVIEW that may result in information disclosure or arbitrary code execution.  Successful exploitation requires an attacker to get a user to open a specially crafted VI.  This vulnerability affects NI LabVIEW 2026 Q3 (26.3.0) and prior versions.

OpenEXR is the reference implementation and specification for the EXR image format, widely used in the motion picture industry. In OpenEXRUtil versions 3.3.0 through 3.3.12 and 3.4.0 through 3.4.12, the documented TypedDeepImageChannel<T>::row() API can return an out-of-bounds pointer when a deep image has a non-zero dataWindow origin, resulting in a heap out-of-bounds read and crash, with potential information disclosure under a controlled heap layout. The flaw arises because ImfDeepImageChannel uses two conflicting coordinate models: at(x, y) uses absolute coordinates (with _base offset by dataWindow.min), while row(r) is documented as 0-based logical access. For a non-zero dataWindow.min, row(0) therefore points outside the _sampleListPointers allocation instead of at the first logical row. This issue is fixed in versions 3.3.13 and 3.4.13.

There is a memory corruption vulnerability recently discovered in NI LabVIEW that may result in information disclosure or arbitrary code execution.  Successful exploitation requires an attacker to get a user to open a specially crafted VI.  This vulnerability affects NI LabVIEW 2026 Q3 (26.3.0) and prior versions.

There is a memory corruption vulnerability recently discovered in NI LabVIEW that may result in information disclosure or arbitrary code execution.  Successful exploitation requires an attacker to get a user to open a specially crafted VI.  This vulnerability affects NI LabVIEW 2026 Q3 (26.3.0) and prior versions.

There is a memory corruption vulnerability recently discovered in NI LabVIEW that may result in information disclosure or arbitrary code execution.  Successful exploitation requires an attacker to get a user to open a specially crafted VI.  This vulnerability affects NI LabVIEW 2026 Q3 (26.3.0) and prior versions.

OpenEXR is the reference implementation and specification for the EXR image format, widely used in the motion picture industry. OpenEXR versions 3.3.0 through 3.3.12 and 3.4.0 through 3.4.13 are vulnerable to a heap out-of-bounds write when exrmetrics reads a crafted deep scanline EXR. This occurs with pixel conversion options such as --pixelmode float or --bench because DeepSlice requests FLOAT output while the backing sample buffers are allocated using the input HALF element size. The issue is fixed in versions 3.3.13 and 3.4.14.

Webkul QloApps does not validate request parameters before a database query. A remote, authenticated attacker with administrative privileges could send a crafted SQL query to the 'bo_query' parameter in the 'Address.php' file. Fixed in 123c97c.

Webkul QloApps does not validate request parameters before a database query. A remote, authenticated attacker with administrative privileges could send a crafted SQL query to the 'bo_query' parameter in the 'CustomerMessage.php' file. Fixed in 123c97c.

There is a memory corruption vulnerability recently discovered in NI LabVIEW that may result in information disclosure or arbitrary code execution.  Successful exploitation requires an attacker to get a user to open a specially crafted VI.  This vulnerability affects NI LabVIEW 2026 Q3 (26.3.0) and prior versions.

Webkul QloApps does not perform proper validation on uploaded file extensions or MIME types before moving the file to a publicly accessible directory. A remote, authenticated attacker with administrative privileges could upload executable files and achieve remote code execution. Fixed in 153ec1c.

OpenEXR is the reference implementation and specification for the EXR image format, widely used in the motion picture industry. In versions before 3.2.11, 3.3.0 through 3.3.12, and 3.4.0 through 3.4.13, a crafted tiled EXR can trigger a heap out-of-bounds write on 32-bit/ILP32 builds when read through the public TiledRgbaInputFile RGBA API. The file uses a small 40x40 dataWindow but a 65537x65537 tile size. On ILP32, the Array2D<Rgba> tile-conversion buffer size calculation overflows, allocates a much smaller heap buffer, and tile decode writes past that allocation. This issue is fixed in versions 3.2.11, 3.3.13, and 3.4.14.

OpenEXR is the reference implementation and specification for the EXR image format, widely used in the motion picture industry. Versions before 3.2.11, 3.3.0 through 3.3.12, and 3.4.0 through 3.4.13 allow a crafted EXR with a nonzero dataWindow.min to make TypedFlatImageChannel::row() return an invalid heap pointer, causing out-of-bounds or use-after-free writes. This occurs when an application writes rows through FlatHalfChannel::row(). Affected consumers are tools, converters, render pipeline components, or image-processing services that accept untrusted EXR files and use FlatHalfChannel::row() on loaded images. This issue is fixed in versions 3.2.11, 3.3.13, and 3.4.14.

There is a memory corruption vulnerability recently discovered in NI LabVIEW that may result in information disclosure or arbitrary code execution.  Successful exploitation requires an attacker to get a user to open a specially crafted VI.  This vulnerability affects NI LabVIEW 2026 Q3 (26.3.0) and prior versions.

browse-mcp is a Playwright-based headless-browser MCP server for MCP-capable agents. Prior to 0.8.2, browser_download writes a fetched response body to join(save_dir, filename) without validating the caller-controlled save_dir, while browser_save_state and browser_load_state honor a caller-controlled path unchanged. A malicious MCP client, or an autonomous agent steered by indirect prompt injection on a visited page, can choose an arbitrary save_dir or state path and a URL whose response body becomes attacker-controlled file contents, allowing writes to any path the process can reach, including ~/.bashrc, autostart entries, or cron files, and potentially leading to host code execution. The force_fetch fallback also uses a raw fetch() that bypasses the BROWSE_MCP_ALLOWED_ORIGINS origin fence. This issue is fixed in version 0.8.2.

QWED is open-source AI verification infrastructure for deterministic verification of LLM outputs, tool calls, code, schemas, and agent state before production execution. Prior to 5.1.2, the qwed package passes caller-controlled math expressions directly to SymPy parse_expr() without restricted global_dict and local_dict namespaces, allowing Python eval() to resolve builtins and execute arbitrary Python code in the API server process. In src/qwed_new/api/main.py, POST /verify/math is protected by get_current_tenant but accepts any valid tenant API key, reads the expression field, applies only a cosmetic re.sub(r'(\d)(()', r'\1*\2', expression) normalization, and passes the result to parse_expr(). In src/qwed_new/core/batch.py, POST /verify/batch sends math items through batch_service.create_job(), stores item.query verbatim, and _verify_item() passes VerificationType.MATH input to parse_expr() without sanitization. The default-enabled POST /auth/signup endpoint allows anyone to create a standard tenant account, POST /auth/api-keys issues an x-api-key, and either vulnerable path can then be used to read or write files, modify data, execute operating system commands, terminate the service, and compromise other tenants in a shared deployment. This issue is fixed in version 5.1.2.

** REJECT ** DO NOT USE THIS CANDIDATE NUMBER. ConsultIDs: CVE-2026-78270. Reason: This candidate is a reservation duplicate of CVE-2026-78270. Notes: All CVE users should reference CVE-2026-78270 instead of this candidate. All references and descriptions in this candidate have been removed to prevent accidental usage.

The Kaltura HTML5 player (mwEmbed / html5lib) contains a local file disclosure vulnerability due to improper validation of the ServiceUrl parameter in mwEmbedLoader.php. This parameter is used as the base URL for a backend request and accepts non‑HTTP schemes such as file://. When an exception or error occurs, the response is subsequently deserialized and its raw contents are reflected to the client in an error message; this enables an unauthenticated, remote attacker to read any arbitrary internal file reachable by the server. Affected versions include html5lib v2.45, v2.103 and earlier, and other v2.x releases exposing the vulnerable endpoint.

The Kaltura HTML5 player (mwEmbed / html5lib) contains an unauthenticated remote code execution vulnerability caused by unsafe data deserialization and unsanitized filesystem path construction. mwEmbedLoader.php accepts a user‑controlled ServiceUrl, whose response is passed to unserialize(), and the resulting object’s fields are written to a cache path derived from attacker‑supplied uiconf_id without proper path validation. An attacker can write arbitrary files into web‑accessible locations and achieve code execution as the webserver user. Affected versions include html5lib v2.45, v2.103 and earlier, and other v2.x releases exposing the vulnerable endpoint.

urllib is an HTTP client for Node.js that supports authentication, redirects, timeouts, and other request features. Prior to 4.9.1 and 2.44.1, urllib follows redirects through followRedirect but reuses caller-supplied options across origins. In src/HttpClient.ts, #requestInternal recursively calls this.#requestInternal(nextUrl.href, options, requestContext), causing options.headers and auth or digestAuth values to be reused when the redirect target has a different scheme, host, or port. Authorization, Cookie, Proxy-Authorization, x-api-key, x-auth-token, and x-access-token can therefore be sent to an attacker-controlled redirected origin, exposing credentials intended for the original origin and potentially allowing reuse against the original partner API or related services. No user interaction is required. This issue is fixed in versions 2.44.1 and 4.9.1.

NVIDIA UFM Enterprise contains a vulnerability in the session management component, where an attacker could use a hard-coded cryptographic key to extract information. A successful exploit of this vulnerability might lead to information disclosure and escalation of privileges.

NVIDIA UFM Enterprise contains a vulnerability in the IBDiagnet API where an authenticated attacker with administrative privileges may cause command injection by sending crafted API requests. A successful exploit of this vulnerability may lead to code execution, escalation of privileges and information disclosure.

NVIDIA UFM Enterprise contains a vulnerability in the user management component, where an authenticated administrator could inject commands by sending a crafted API request. A successful exploit of this vulnerability might lead to code execution, escalation of privileges and information disclosure.

NVIDIA UFM Enterprise contains a vulnerability in the plugin management API, where an authenticated user with low privileges could inject code by sending a specially crafted API request. A successful exploit of this vulnerability might lead to code execution, escalation of privileges and information disclosure.

NVIDIA UFM Enterprise contains a vulnerability in the web interface authorization component, where an authenticated user could cause improper authentication by sending specially crafted HTTP requests. A successful exploit of this vulnerability might lead to code execution and escalation of privileges.

There is an integer overflow vulnerability resulting in an out-of-bounds write recently discovered in NI LabVIEW.  This may result in information disclosure or arbitrary code execution.  Successful exploitation requires an attacker to get a user to open a specially crafted VI file.  This vulnerability affects NI LabVIEW 2026 Q3 and prior versions.

NVIDIA DGX Spark contains a vulnerability in UEFI where a Attacker may cause a/an CWE-693 by privileged local user. A successful exploit of this vulnerability may allow an attacker to bypass administrator password protection in UEFi.

NVIDIA DGX Spark contains a vulnerability in the standalone MM firmware where an attacker could be able to cause an out-of-bounds read. A successful exploit of this vulnerability might lead to information disclosure.

NVIDIA DGX Spark contains a vulnerability in the system firmware, where a privileged attacker could be able to cause a NULL pointer dereference. A successful exploit of this vulnerability may lead to code execution, escalation of privileges, denial of service, information disclosure, and data tampering.

NVIDIA DGX Spark contains a vulnerability in the system firmware, where a privileged attacker could be able to cause an out-of-bounds write. A successful exploit of this vulnerability may lead to code execution, escalation of privileges, denial of service, information disclosure, and data tampering.

NVIDIA DGX Spark contains a vulnerability in the system firmware, where a privileged attacker could be able to cause an out-of-bounds write. A successful exploit of this vulnerability may lead to code execution, escalation of privileges, denial of service, information disclosure, and data tampering.

There is an integer conversion vulnerability resulting in an out-of-bounds read when loading images recently discovered in NI LabVIEW.  This may result in information disclosure or arbitrary code execution.  Successful exploitation requires an attacker to get a user to open a specially crafted VI file.  This vulnerability affects NI LabVIEW 2026 Q3 and prior versions.

djust provides Phoenix LiveView-style reactive server-side rendering for Django with Rust-powered performance. Prior to 1.0.4, LiveViewConsumer.handle_mount sends a `{"type":"navigate","to":...}` frame when login_required, permission_required, or a redirecting on_mount hook denies a LiveView mount, but returns without closing the WebSocket or clearing self.view_instance. A browser follows the redirect, but a raw WebSocket client can ignore it and retain the mounted socket. Because LiveViewConsumer.handle_event does not recheck authentication or authorization, the client can send `{"type":"event",...}` frames that invoke @event_handler methods without an authenticated session, including through handle_live_redirect_mount, enabling unauthorized sensitive reads or mutations. This issue is fixed in version 1.0.4.

The Zephyr ext2 filesystem driver validates the on-disk block bitmap in ext2_init_fs() (subsys/fs/ext2/ext2_impl.c) by passing fs_blocks = s_blocks_count - s_first_data_block to ext2_bitmap_count_set(). That helper (subsys/fs/ext2/ext2_bitmap.c) treats its argument as a number of bits and reads one bitmap byte per eight bits, but the bitmap buffer (BGROUP_BLOCK_BITMAP) is a single fetched block of only fs->block_size bytes (capacity fs->block_size * 8 bits). s_blocks_count and s_first_data_block are taken verbatim from the superblock and were never bounded against this single-group capacity; ext2_verify_disk_superblock() checks the magic, revision, and block-size shift but not the block count. A crafted ext2 image with an oversized s_blocks_count (up to ~4 billion, against a maximum 4096-byte block / 32768-bit bitmap) makes ext2_bitmap_count_set() scan roughly 512 MB of memory past the bitmap block — a large out-of-bounds read of the static block slab and adjacent memory. The defect is reached during mount: ext2_init_fs() is invoked from ext2_mount() (subsys/fs/ext2/ext2_ops.c), the registered .mount operation. Any path that mounts an attacker-supplied ext2 image (removable media, a disk/flash partition, or a downloaded image) triggers it. The kernel-privileged parser operates on attacker-controlled data, so the bug is exploitable wherever untrusted ext2 media can be mounted. Impact is an out-of-bounds read only: the resulting bit count is compared internally and the mount is rejected, so no attacker-controlled bytes are returned (not a useful information leak). The ~512 MB over-read will almost certainly cross an unmapped or MPU-protected boundary and fault, crashing the system — a denial of service triggered by mounting a single malformed image. The fix rejects any image whose fs_blocks exceeds fs->block_size * 8 before the scan.

The OCPP 1.6 client in subsys/net/lib/ocpp/ocpp.c reconstructs a session handle and PDU id from the uid field of a CALLRESULT message. In ocpp_process_server_msg() the code calls atoi(strtok_r(uid, "-", &tmp)) without checking the strtok_r return value. When the server-supplied uid is empty or contains no - delimiter, strtok_r() returns NULL and atoi(NULL) dereferences a NULL pointer, which is undefined behaviour. The uid originates from network data: parse_rpc_msg() in subsys/net/lib/ocpp/ocpp_j.c JSON-parses a frame received from the OCPP central system over TCP/WebSocket and copies the server-controlled string into the local buffer. A malicious or compromised central system, or a man-in-the-middle on a non-TLS ws:// connection, can return a malformed uid to reach the defect. No authentication beyond the existing server connection (or MITM position) is required, and the reconstructed pointer is membership-validated by ocpp_session_is_valid(), so the impact is limited to the NULL dereference rather than arbitrary pointer use. On Zephyr targets that trap access to address 0 (MMU/MPU platforms or CONFIG_NULL_POINTER_EXCEPTION_DETECTION), the dereference faults inside the OCPP reader thread and invokes the fatal handler, producing a remote denial of service of the charge point; on bare targets where address 0 is readable the call returns 0 and is benign, so the impact is availability-only and platform-conditional. The applied fix guards only the first atoi(); the second strtok_r(NULL, "-", &tmp) followed by pdu = atoi(buf) in the same function remains unguarded and the identical NULL dereference is still reachable from the same network input when the uid has a first token but no second --delimited token. A complete fix should validate the second token as well.

The virtio PCI driver (drivers/virtio/virtio_pci.c) parses a device's PCI capability list during driver initialization. In virtio_pci_read_cap() the device-supplied capability length byte cap_len (read from PCI config space via pcie_conf_read()) was only checked with assert(tmp.cap_len == cap_struct_size). That assert resolves to __ASSERT_NO_MSG(), gated by CONFIG_ASSERT, which defaults off in production builds, so the value reached the copy logic completely unvalidated. The length then drives a loop that copies extra capability dwords into a fixed-size stack buffer supplied by the caller. A cap_len below the 24-byte base struct virtio_pci_cap underflows the unsigned extra_data_words count to a near-SIZE_MAX value, producing an effectively unbounded stack write; a cap_len above the caller's buffer (up to 255) writes up to roughly 228 bytes of device-controlled data past the buffer. Both are out-of-bounds writes of attacker-controlled content executed in kernel mode during boot-time device probe. The input originates from the virtio device. In the common deployment where Zephyr runs as a guest under a hypervisor, the device backend is the host, which already fully outranks the guest, so the bug yields no privilege escalation. The exploitable case is a virtio device that is untrusted relative to the Zephyr kernel — an untrusted or physical/passthrough virtio PCIe device on a bare-metal system, or a confidential-computing posture where the guest must defend against the host — where a malicious device can corrupt the kernel stack and potentially achieve code execution or a crash. The fix replaces the compiled-out assert with a runtime range check rejecting cap_len outside [sizeof(struct virtio_pci_cap), cap_struct_size] before any arithmetic or copy.

Nextcloud MCP Server is a production-ready MCP server that connects AI assistants to a Nextcloud instance. Prior to 0.117.2, the POST /webhooks/nextcloud endpoint in nextcloud_mcp_server/vector/webhook_receiver.py has no authentication by default because WEBHOOK_SECRET defaults to None and startup validation does not require it. When WEBHOOK_SECRET is unset, handle_nextcloud_webhook() accepts unauthenticated requests. The payload["user"]["uid"] field parsed in nextcloud_mcp_server/vector/webhook_parser.py is attacker-controlled and is used without an authenticated-session cross-check for Qdrant operations, allowing a network attacker to delete or trigger re-indexing of vector embeddings for any user and to destroy the semantic search index by sending forged deletion events. This issue is fixed in version 0.117.2.

X-AnyLabeling's model downloader disabled TLS certificate verification. download_with_retry in anylabeling/services/auto_labeling/model.py built a context with ssl._create_unverified_context() and passed it to urllib.request.urlopen, so neither the certificate chain nor the hostname was checked on any model download, and models are fetched over HTTPS from the project's release host. Any party positioned to intercept that connection could therefore answer it with content of their own choosing. The response is written to a .part file and moved into place with os.replace, and the only post-download check, safe_check_model, validates the file's format rather than its provenance: no hash or signature is compared against an expected value. For an ONNX target the substituted file passes onnx.checker.check_model and is then used for inference, so the attacker chooses the model that produces the application's annotations. For a .pth or .pt target, which the shipped SAM2 video, YOLOE, UPN and open_vision configurations use, the check worker calls torch.load without weights_only, so a substituted file is unpickled and executes code of the attacker's choosing on PyTorch releases predating the weights_only default.

mcp-shell is an MCP server for running shell commands securely, auditably, and on demand. Prior to 0.6.0, config.go initializes Security.Enabled to false, and when MCP_SHELL_SEC_CONFIG_FILE is unset, main.go starts the documented bare-binary deployment without a security policy. SecurityValidator.validateCommand in security.go then short-circuits and allows every command supplied to the shell_exec MCP tool, so an LLM connected over stdio can execute unrestricted OS commands as the mcp-shell process user. The README from-source installation and MCP client configuration omit MCP_SHELL_SEC_CONFIG_FILE, making the insecure state the documented default. This issue is fixed in version 0.6.0.

mcp-shell is an MCP server for running shell commands securely, auditably, and on demand. Prior to 0.6.0, the default Docker security.yaml includes /bin/bash in allowed_executables, while security.go validates only the first token and checkBlockedPatternsAndCommands does not reject the shell command-mode flag -c. A caller of the shell_exec MCP tool can provide the command argument `/bin/bash -c <arbitrary-command>`, which passes validation and reaches executor.go, where parseCommand and exec.CommandContext execute the arbitrary command as mcpuser outside the intended allowlist. This issue is fixed in version 0.6.0.

mcp-shell is an MCP server for running shell commands securely, auditably, and on demand. Prior to 0.6.0, the default security.yaml allows /usr/bin/git, while security.go omits ! from containsShellMetacharacters and containsDangerousShellConstructs and applies no per-executable argument policy. A caller of the shell_exec MCP tool can provide the command argument /usr/bin/git -c alias.pwn=!<arbitrary-command>, causing Git to create a shell alias and execute arbitrary OS commands as the mcp-shell process user. The default Docker image runs as mcpuser with Git installed and secure mode enabled, so the bypass is exploitable in the default deployment without additional authentication beyond MCP connectivity. This issue is fixed in version 0.6.0.

QWED-MCP is a deterministic verification gateway for MCP. Prior to 0.2.1, verify_math_expression() in src/qwed_mcp/engines/math_engine.py passes attacker-controlled expression and claimed_result strings directly to SymPy's parse_expr() after only normalizing caret syntax to Python exponent syntax, without restricting global_dict, removing Python built-ins, or validating the expression AST. Because parse_expr() calls Python's eval() with built-ins available, an attacker who can cause a downstream caller to pass untrusted input to this public library function can use Python import functionality to execute arbitrary operating-system commands as the qwed-mcp process user, read or modify accessible data, exfiltrate process secrets, or reach internal services. The default MCP tool registry does not expose verify_math_expression(), so exploitation requires a downstream integration that invokes the library API with attacker-controlled input. This issue is fixed in version 0.2.1.

An authorization weakness in JFrog Artifactory Composer repository handling may allow an authenticated user, under specific conditions, to read package metadata from repositories they are not authorized to read. The issue affects confidentiality and has been addressed in fixed Artifactory versions.

PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.58, Browser Server _handle_connection() checks Chrome extension origins with re.match() and the unanchored expression chrome-extension://[a-z0-9]{32}. Extra trailing characters pass before websocket.accept(), allowing start_session commands and unauthorized browser automation. This issue is fixed in version 4.6.58.

PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.58, MCP HTTP Stream _validate_origin uses request_origin.startswith(allowed), allowing the attacker-controlled localhost.attacker.com HTTP origin to satisfy the localhost allowlist. A webpage can send Content-Type: text/plain requests without preflight and invoke tools/call without an API key, including file writes that persist agent instructions. This issue is fixed in version 4.6.58.

Under specific circumstances, low-level user can run request to remote CocoaPods repos via JFrog Artifactory External Dependency.

Vocos instantiates a class named by a configuration file without restricting which class may be named. instantiate_class in vocos/pretrained.py takes the class_path value from the configuration, splits it into a module and an attribute, imports the module with __import__, resolves the attribute with getattr, and calls the result as args_class(*args, **kwargs) where kwargs is the config's own init_args mapping. No allowlist constrains the dotted path, so a configuration may name any importable callable and supply the arguments it is called with. Vocos.from_hparams reaches this for each of the feature_extractor, backbone and head entries, and Vocos.from_pretrained reaches it with a remote file: it downloads config.yaml from a caller-named Hugging Face repository and passes it straight to from_hparams. Loading a model from a repository the user does not control therefore executes code of the repository owner's choosing in the loading process. The neighbouring torch.load of the downloaded weights is a separate matter and is constrained on PyTorch releases that default weights_only to true, which leaves this path as the reachable one.

rclone before 1.74.4 fails to strip the X-Amz-Security-Token header when an S3 redirect changes scheme from HTTPS to HTTP on the same host. Attackers can intercept plaintext HTTP traffic to capture AWS STS session tokens sent in request headers.

rclone before 1.74.4 fails to mask special permission bits when applying source-supplied mode metadata in the local backend, allowing attackers to set setuid/setgid bits on attacker-controlled files. When copying with metadata preservation from an untrusted remote, attackers can plant a setuid binary that escalates privileges to root if rclone runs as root, or to the service account user otherwise.

rclone serve s3 before 1.74.4 contains a path traversal vulnerability that allows attackers to read and overwrite root-level files by using dot-dot segments in S3 object keys. Attackers can send requests with object keys like ../root-secret.txt to escape the bucket namespace and access files in the serve root directory.

rclone versions before v1.75.0 fail to reject transport downgrades in redirect handling, allowing Basic authorization and Cookie headers to be replayed over plaintext HTTP after same-host HTTPS-to-HTTP redirects. An on-path attacker observing the plaintext hop can capture and reuse credentials to perform WebDAV operations with the compromised account's permissions.