CVE-20XX-?
agent?

Just published.

The newest CVEs, straight from CVE.org and NVD.

410,160 CVEs · updated 2026-11-21 02:07 UTC

Reset all
  1. CVE-2026-90591

    The NXP ENET-QOS Ethernet driver's receive work handler eth_nxp_enet_qos_rx() in drivers/ethernet/eth_nxp_enet_qos hands each processed descriptor back to the DMA by writing desc->read.control = rx_desc_refresh_flags, which sets OWN_FLAG and BUF1_ADDR_VALID_FLAG. RDES0 of a receive descriptor is the DMA buffer address in read format, but the hardware write-back aliases that word with the VLAN tag fields (or, for a context descriptor, the timestamp low word), so it is destroyed on every completion. The success path rewrites desc->read.buf1_addr before re-arming; the four drop paths — errored or non-first descriptor, net_pkt_rx_alloc() failure, invalid descriptor length, and net_pkt_get_frag() failure — did not, re-arming the ring slot with the write-back residue as its DMA target address. The drop paths are driven entirely by received traffic and need no privilege or special configuration. With the defaults for this driver (CONFIG_NET_BUF_DATA_SIZE of 128 and 16 receive descriptors) a single full-size frame consumes twelve ring slots and twelve K_NO_WAIT fragment allocations, so a short burst of frames sent to the device exhausts the receive pools and reaches the allocation-failure paths; a frame with a receive error, or any frame whose first descriptor was already dropped, reaches the non-first-descriptor path. Once a slot has been poisoned, the DMA writes the next frame received in that slot to an address derived from the frame's VLAN tag fields rather than to a receive buffer. On the supported MCX SoCs that address falls in flash or reserved space, so the transfer raises a fatal bus error that the driver only logs — there is no reset or recovery path — and reception stops permanently until the device is rebooted; where the write is silently discarded instead, the untouched buffer is later handed to the network stack with a length, delivering stale buffer contents as if they were a received frame. A single triggering event is enough to make the condition persistent, so an attacker able to send traffic to the device can permanently disable its network interface.

    Memory safety OS & kernel

    7.5 HIGH 0.3% exploit odds Fix listed
  2. CVE-2026-90585

    The Renesas RA SCI UART driver's receive-error ISR, uart_ra_sci_eri_isr() in drivers/serial/uart_renesas_ra_sci.c, acknowledged the error condition only when a Zephyr interrupt-driven callback had already been registered. With CONFIG_UART_INTERRUPT_DRIVEN=y and CONFIG_UART_ASYNC_API=n, if data->user_cb is NULL the ISR body is empty: the SCI error flags (SSR.ORER, SSR.FER, SSR.PER) are left set and the ICU event flag R_ICU->IELSR_b[eri_irq].IR is never cleared, so the interrupt re-pends immediately and the ISR is re-entered without end. The error interrupt is masked at boot by uart_ra_sci_init(), but FSP's R_SCI_UART_Open() unconditionally re-enables it, and uart_ra_sci_configure() - the back end of the public uart_configure() API, enabled by default via CONFIG_UART_USE_RUNTIME_CONFIGURE - performs a Close()/Open() pair without re-masking it. Any runtime uart_configure() therefore leaves the error interrupt live regardless of whether a callback has been set, as does calling uart_irq_err_enable() on its own. An instance used with the polling API, or the window between uart_configure() and uart_irq_callback_user_data_set(), thus runs with the error interrupt unmasked and no callback registered. The trigger is an ordinary receive line error - overrun, framing, parity, or a break condition - which is fully determined by the device attached to the serial link. A peer able to assert a break or transmit at a mismatched baud rate produces such an error at will, so a single frame is enough to enter the state. The impact is availability only: the CPU is captured in the re-entering ISR and never returns to thread context, hanging the entire system until an external reset or hardware watchdog fires. There is no memory corruption or information disclosure. Builds using the asynchronous UART API are unaffected, because the fall-through to FSP's sci_uart_eri_isr() clears both the error flags and the pending IRQ.

    Denial of service OS & kernel

    4.6 MEDIUM 0.2% exploit odds Fix listed
  3. CVE-2026-90586

    The Xtensa MMU userspace port never initialized the per-thread page-table pointer thread->arch.ptables during thread setup, while arch_mem_domain_thread_add() in arch/xtensa/core/ptables.c uses a non-NULL value of that field as its "this thread is migrating from another memory domain" predicate. For a thread whose struct k_thread backing memory is not zeroed, the field holds stale data and the kernel takes the migration branch, passing that stale value to reset_region() as the base of a level-1 page table. The equivalent field on x86 is explicitly cleared in z_x86_userspace_prepare_thread() for exactly this reason; the Xtensa port adopted the predicate without the initialization. The condition is reachable from an unprivileged user-mode thread on a build with CONFIG_XTENSA_MMU, CONFIG_USERSPACE and dynamically allocated kernel objects. k_object_alloc(K_OBJ_THREAD) is a system call and allocates the thread object from the caller's resource pool without zeroing it, and the k_thread_create() verifier accepts such an object while forcing the K_USER option, so the migration branch is taken whenever the recycled heap bytes at that offset are non-zero. Because a user thread can also allocate, fill and free other kernel objects from the same heap, the stale bytes are groomable rather than merely random. Walking the bogus table in region_map_update() performs an unguarded index into l2_page_tables_counter[], may copy from and write through the attacker-influenceable level-1 base in dup_l2_table_if_needed(), and finally writes a page-table entry through a second pointer read out of that memory — once per page of the new thread's stack. The result is out-of-bounds reads and writes executed in supervisor mode, ranging from an immediate kernel panic (observed as Xtensa EXCAUSE 13 on Intel ADSP DSP firmware with user threads) to corruption of kernel memory and escape from the user-mode sandbox.

    Memory safety OS & kernel

    7.8 HIGH 0.1% exploit odds Fix listed
  4. CVE-2026-90587

    usbh_device_init() in subsys/usb/host/usbh_device.c rejects a device whose device descriptor advertises bNumConfigurations == 0, but it only logs the error and jumps to the cleanup label without setting the return code. At that point the local err is still 0 from the preceding successful descriptor read, so the function returns success while having skipped usbh_device_set_configuration() — the only place udev->cfg_desc is allocated. The device structure is zero-initialised by usbh_device_alloc(), so udev->cfg_desc remains NULL. usbh_device_connect() treats the zero return as a successful enumeration and calls usbh_class_probe_device(), which starts its descriptor walk from udev->cfg_desc. usbh_desc_get_next_function() passes that NULL pointer to usbh_desc_is_valid(), whose only guard compares the caller-supplied size, so head->bLength is read through a NULL pointer. The condition is triggered entirely by the attached peripheral: attaching a USB device (or any USB gadget emulating one) that reports bNumConfigurations = 0 reaches the defect directly from the UHC_EVT_DEV_CONNECTED_FS/_HS event handler in subsys/usb/host/usbh_core.c, with no authentication and no user interaction beyond the physical attachment. On targets where page zero is unmapped or CONFIG_NULL_POINTER_EXCEPTION_DETECTION is enabled this raises a fatal exception in the USB host thread, denying service to the host. Where address zero is readable, the descriptor walker instead parses low memory as a descriptor chain and can dispatch a class probe() with attacker-influenced values, causing further out-of-bounds reads through the still-NULL udev->cfg_desc in usbh_desc_get_iface(). No attacker-controlled write was identified, so the impact is availability plus a bounded out-of-bounds read. Note that CONFIG_USB_HOST_STACK is marked experimental.

    Memory safety

    4.6 MEDIUM 0.2% exploit odds Fix listed
  5. CVE-2026-84410

    In apport before 2.37.0, an improper privilege management vulnerability allows an unprivileged local attacker to gain root privileges on systems where automatic crash reporting is enabled. The automatic crash report processing pipeline runs as root and executes ldd on the attacker-controlled ExecutablePath field of a crash file placed in the world-writable /var/crash directory, causing the dynamic loader to load and execute an attacker-supplied audit module specified by the ELF's DT_AUDIT tag. This allows an attacker to execute arbitrary code with root privileges and fully compromise the affected system.

    Auth & access Library

    7.3 HIGH 0.1% exploit odds Fix listed
  6. CVE-2026-86692

    In apport before 2.37.0, a path traversal vulnerability allows an attacker who can supply a crafted crash report to cause unintended file creation or overwrite outside the intended extraction directory when the report is processed with apport-unpack. The key-name validation added for CVE-2026-77113 does not cover base64-encoded keys skipped by ProblemReport.load(binary=False), and ProblemReport.extract_keys() writes them to a file path built from the key without validation. This allows an attacker to create or overwrite arbitrary files writable by the invoking user, such as .bashrc or .profile, gaining persistence or code execution in that user's context.

    Disclosure

    5.7 MEDIUM 0.2% exploit odds Fix listed
  7. CVE-2026-73857

    ModSecurity is an open source, cross platform web application firewall (WAF) engine for Apache, IIS and Nginx. From 3.0.15 until 3.0.17, XML request body processing with SecParseXmlIntoArgs set to On or OnlyArgs can leave XMLNodes::parsing_ctx_arg uninitialized in the SAX end-element callback. A single crafted unauthenticated HTTP request can make a worker dereference the pointer, resulting in a constant-value write to an attacker-controlled address. The resulting memory corruption can terminate the worker and may disclose memory or redirect execution. Earlier 3.0.x releases and ModSecurity 2.x do not use the affected path. This issue is fixed in version 3.0.17.

    Memory safety Network gear

    7.5 HIGH 0.4% exploit odds Fix listed
  8. CVE-2026-77473

    AsyncSSH is a Python package which provides an asynchronous client and server implementation of the SSHv2 protocol on top of the Python asyncio framework. Prior to 2.24.1, recursive remote-to-local SFTP copies in SFTPClient._copy() in asyncssh/sftp.py, reached through get(), mget(), and copy(), trust filenames returned by a remote server's directory enumeration. A malicious or compromised SFTP server can return an absolute filename or a filename containing path separators and parent-directory traversal, causing destination path construction to escape the requested local directory and write attacker-selected content to locations writable by the client process. This issue is fixed in version 2.24.1.

    Disclosure Library

    6.5 MEDIUM 0.5% exploit odds Fix listed
  9. CVE-2026-63497

    titra is an open source time tracking project. In versions 1.0.3 and prior, authenticated API users can abuse the customfields object to inject or overwrite top-level fields in MongoDB documents. At time of publication, there are no publicly known patches.

    Injection

    5.3 MEDIUM 0.2% exploit odds Fix listed
  10. CVE-2026-63496

    augment-vir is a collection of helper functions. Prior to version 31.73.5, the grep() function exported by @augment-vir/node inserts caller-controlled search patterns directly into a double-quoted shell command string without sanitization (grep.ts:310). When the constructed command is passed to spawn() with shell: 'bash' (shell.ts:98-99), bash evaluates $() command substitution expressions embedded in the pattern, resulting in arbitrary OS command execution as the Node.js process user. Any downstream application that passes untrusted input to grep()'s pattern or patterns field is vulnerable to unauthenticated local (and potentially remote) code execution. This issue has been patched in version 31.73.5.

    Injection Device

    7.8 HIGH 0.2% exploit odds Fix listed
  11. CVE-2026-63634

    pympp is a Python SDK for the Machine Payments Protocol. Prior to version 0.9.1, when the server acts as the fee_payer, pympp validates calldata using eth_abi decoding, which is lenient about trailing bytes. The FeePayerPolicy caps gas_limit (≤ 2 M) and max_fee_per_gas (≤ 100 Gwei) but does not check calldata length. Tempo uses legacy calldata gas pricing: 16 gas per non-zero byte. Appending N bytes of 0x01 padding inflates actual gas used by N × 16 while keeping gas_limit and max_fee_per_gas within policy caps. The server cosigns and broadcasts the padded transaction, draining its fee-payer wallet. This issue has been patched in version 0.9.1.

    Other Library

    6.9 MEDIUM 0.3% exploit odds Fix listed
  12. CVE-2026-63644

    bird-lg-go is a BIRD looking glass in Go. Prior to version 1.4.7, unsanitized input path enables arbitrary local file read, allowing an unauthenticated remote attacker to read sensitive files such as /etc/shadow and /root/.bash_history. This issue has been patched in version 1.4.7.

    Injection

    7.5 HIGH 0.4% exploit odds Fix listed
  13. CVE-2026-61812

    ModSecurity is an open source, cross platform web application firewall (WAF) engine for Apache, IIS and Nginx. Prior to 2.9.15 and 3.0.17, t:htmlEntityDecode recognizes only a small subset of named HTML entities, so rules that depend on the transformation can inspect different bytes than a browser or backend. Attackers can encode ASCII separators and operators with unhandled entity names, including aliases, to bypass rules that depend on the transformation. The v3 decoder also accepts prefix matches instead of requiring an exact entity-token length, which can transform invalid longer names into different bytes. This issue is fixed in versions 2.9.15 and 3.0.17.

    Other Browser & mail

    6.3 MEDIUM 0.3% exploit odds Fix listed
  14. CVE-2026-56712

    Replicate Cog through 0.23.1 contains a server-side request forgery vulnerability that allows attackers to cause the Cog process to make server-side HTTP requests by submitting prediction inputs with attacker-controlled URLs that are not validated against the resolved destination host or IP before fetching. Attackers can supply malicious URLs or exploit HTTP redirects in cog.File and cog.Path predictor inputs to force requests to internal addresses, potentially exposing sensitive internal services or credentials.

    Disclosure Server

    7.8 HIGH 0.5% exploit odds Fix listed
  15. CVE-2026-125109

    Rack is a modular Ruby web server interface. From 3.0 until 3.1.23 and 3.2.8, Rack::Request#host_with_port, #base_url, and #url can return an attacker-controlled invalid authority from Host, X-Forwarded-Host, or Forwarded: host= verbatim. Applications without an independent host allowlist that use these methods for absolute URLs, redirects, or origin-dependent checks can emit attacker-chosen delimiters, enabling redirect manipulation, string-based allowlist bypass, or unescaped markup reflection. Rails overrides these methods and is not expected to be affected. This issue is fixed in versions 3.1.23 and 3.2.8.

    Other Server

    4.8 MEDIUM 0.2% exploit odds Fix listed
  16. CVE-2026-125110

    Rack is a modular Ruby web server interface. Prior to 2.2.25, 3.1.23, and 3.2.8, Rack::Static serves a path after Utils.clean_path_info normalization but evaluates header_rules against a path that is only URL-decoded. An unauthenticated attacker can request a known file through raw or percent-encoded dot segments and receive it without path-specific security headers such as Content-Disposition or Content-Security-Policy. Applications that do not use path-specific header_rules are not affected. This issue is fixed in versions 2.2.25, 3.1.23, and 3.2.8.

    Other Server

    5.3 MEDIUM 0.3% exploit odds Fix listed
  17. CVE-2026-125111

    Rack is a modular Ruby web server interface. Prior to 2.2.25, 3.1.23, and 3.2.8, Rack::Multipart::Parser can retain parsed data in its internal StringScanner buffer when reads end inside or immediately after multipart headers. An unauthenticated attacker can align many small parts with rack.multipart.buffer_size boundaries so the buffer grows with the request despite per-field, header, and preamble limits, causing a worker to exhaust memory. The attack requires only a well-formed multipart body and affects applications that parse multipart requests. This issue is fixed in versions 2.2.25, 3.1.23, and 3.2.8.

    Denial of service Library

    7.5 HIGH 0.4% exploit odds Fix listed
  18. CVE-2026-125112

    djust provides Phoenix LiveView-style reactive server-side rendering for Django with Rust-powered performance. Prior to versions 1.0.10 and 1.1.2, djust keys context output-safety grants by variable name and fails to revoke a grant when a template bind replaces the value associated with that name. A template using {% with %}, {% for %}, {% include ... with %}, tuple unpacking, or assign tags can therefore rebind a name that previously held mark_safe content and emit the new attacker-controlled value without escaping, including descendant lookups through the rebound name. This can cause stored or reflected cross-site scripting when a victim renders the affected template. This issue is fixed in versions 1.0.10 and 1.1.2.

    XSS Web app

    2.3 LOW 0.3% exploit odds Fix listed
  19. CVE-2026-125105

    Wagtail is an open source content management system built on Django. Prior to 7.0.9, 7.3.4, 7.4.3, and 8.0rc2, Wagtail applies ETag conditional request processing to the document serve endpoint before document permission and filename checks complete. An unauthenticated attacker who knows a document ID and a candidate SHA1 hash can submit If-Match or If-None-Match and distinguish the response, revealing whether that hash is present in the document library. The issue does not expose document contents, filenames, other metadata, or information about hashes the attacker does not already know. This issue is fixed in versions 7.0.9, 7.3.4, 7.4.3, and 8.0rc2.

    Other Web app

    3.7 LOW 0.3% exploit odds Fix listed
  20. CVE-2026-125106

    Wagtail is an open source content management system built on Django. Prior to 7.0.9, 7.3.4, 7.4.3, and 8.0rc2, a Wagtail CMS user with the submit translations permission can call the Admin API copy for translation action for a source page that the user cannot edit. The action copies the inaccessible page into a translation, allowing the user to view its contents despite the missing source-page edit permission. This issue is fixed in versions 7.0.9, 7.3.4, 7.4.3, and 8.0rc2.

    Other Web app

    6.5 MEDIUM 0.3% exploit odds Fix listed
  21. CVE-2026-125107

    Muhammara is a node module with c/cpp bindings to modify PDF with JavaScript for node or electron. Prior to 6.0.5, Muhammara PDFParser::CreateFilterForStream() in src/deps/PDFWriter/PDFParser.cpp dereferences a null PDFInteger pointer when a PDF stream uses the LZWDecode filter with a DecodeParms dictionary that omits the EarlyChange key. An application that processes an untrusted PDF and reads the affected stream can crash, causing a denial of service. This issue is fixed in version 6.0.5.

    Memory safety Library

    7.5 HIGH 0.3% exploit odds Fix listed
  22. CVE-2026-125108

    Rack is a modular Ruby web server interface. Prior to 2.2.25, 3.1.23, and 3.2.8, Rack::Multipart::Parser.parse_boundary can skip a leading empty or quoted boundary parameter in a multipart Content-Type header and accept a later parameter. Rack can therefore parse a request body with a boundary that a conforming intermediary would not derive, allowing form fields or file uploads to reach an application without the intended proxy or web application firewall inspection. This issue is fixed in versions 2.2.25, 3.1.23, and 3.2.8.

    Other Network gear

    3.7 LOW 0.2% exploit odds Fix listed
  23. CVE-2026-124992

    A vulnerability was identified in nushell up to 0.116.1. Impacted is the function escape_bytes of the file crates/nu-json/src/ser.rs of the component Data Conversion Handler. The manipulation leads to improper neutralization of escape, meta, or control sequences. The attack must be carried out locally. The pull request to fix this issue awaits acceptance.

    Other

    1.9 LOW 0.1% exploit odds Fix listed
  24. CVE-2026-124994

    A security flaw has been discovered in Turboderp-org ExLlamaV3 up to 1.6.3. The affected element is the function ngram_dequant of the file exllamav3/exllamav3_ext/ngram.cu of the component NGramEmbedding. The manipulation of the argument head_bias results in out-of-bounds read. The attack may be launched remotely. The exploit has been released to the public and may be used for attacks. Upgrading to version 1.6.4 is sufficient to fix this issue. The patch is identified as c9a57ae832a3035e264e0a412e1daf4ebb699b98. It is suggested to upgrade the affected component.

    Memory safety

    2.1 LOW 0.3% exploit odds Fix listed
  25. CVE-2026-124996

    A weakness has been identified in Turboderp-org ExLlamaV3 up to 1.6.3. The impacted element is the function routing_sel_norm_kernel of the file exllamav3/exllamav3_ext/routing.cu of the component hash-MoE routing. This manipulation of the argument gate.tid2eid causes out-of-bounds read. Remote exploitation of the attack is possible. Upgrading to version 1.6.4 is sufficient to resolve this issue. Patch name: c9a57ae832a3035e264e0a412e1daf4ebb699b98. Upgrading the affected component is recommended.

    Memory safety

    5.3 MEDIUM 0.2% exploit odds Fix listed
  26. CVE-2026-124999

    A security vulnerability has been detected in braindecode up to 1.8.2. This affects the function EEGModuleMixin._from_pretrained of the file braindecode/models/base.py of the component Hub Integration. Such manipulation of the argument filename leads to path traversal. The attack can be executed remotely. The exploit has been disclosed publicly and may be used. Upgrading to version 2.0.0 mitigates this issue. The name of the patch is 05b764d93b9a50400f523206fbe03d73d0cf5edb. It is recommended to upgrade the affected component.

    Disclosure

    2.1 LOW 0.4% exploit odds Fix listed
  27. CVE-2026-123816

    pyLoad is a free and open-source download manager written in Python. Prior to 0.5.0b3.dev101, Api.check_apikey caches successful validation in self._apikey_cache using only the attacker-controlled key_id. After a legitimate key populates the cache, a forged API key with the same key_id reaches the cache-hit path without comparing apikey[-43:] to the stored hash through hmac.compare_digest. Key identifiers are small sequential integers, so an unauthenticated remote attacker can enumerate an active identifier and inherit the cached key owner's permissions during the 300-second cache lifetime. This issue is fixed in version 0.5.0b3.dev101.

    Auth & access

    8.1 HIGH 0.4% exploit odds Fix listed
  28. CVE-2026-123817

    AIOHTTP is an asynchronous HTTP client/server framework for asyncio and Python. Prior to 3.14.4, the WebSocket reader retains one bytes object for each transport read while reassembling an incomplete frame. A malicious peer can split a frame into many tiny reads so per-object overhead substantially exceeds the payload bytes allowed by max_msg_size. The condition affects a server receiving frames from a malicious client and a client receiving frames from a malicious server, and can cause excessive memory consumption or denial of service. This issue is fixed in version 3.14.4.

    Denial of service Server Aiohttp

    8.7 HIGH 0.4% exploit odds Fix listed
  29. CVE-2026-124033

    The export-cluster-logs and export-image-logs commands in the AWS ParallelCluster CLI before version 3.16.0 contain a relative path traversal issue. An authenticated remote user with write access to the log export Amazon S3 bucket could use a crafted S3 object key to write files outside the intended export directory on the machine running the command. The written files would have names and content controlled by a third party, and would be created with the privileges of the user running the command. To remediate this issue, users should upgrade to version 3.16.0 or later.

    Disclosure

    5.9 MEDIUM 0.4% exploit odds Fix listed
  30. CVE-2026-123818

    LMDeploy is a toolkit for compressing, deploying, and serving large language models. From 0.12.3 until 0.15.0, _is_safe_url validates remote media URLs with urlparse while requests.Session().get interprets backslashes differently when sending the request. An attacker can provide a URL whose validation hostname appears globally routable but whose request destination resolves to an internal host, bypassing non-global address checks. Applications that accept attacker-controlled remote media URLs can therefore be induced to access internal HTTP services and expose data available from those services. This issue is fixed in version 0.15.0.

    Disclosure

    7.5 HIGH 0.3% exploit odds Fix listed
  31. CVE-2026-123819

    Wagtail is an open source content management system built on Django. Prior to 7.0.9, 7.3.4, 7.4.3, and 8.0rc2, a CMS user who has add permission for a snippet model but lacks change and view permission for an existing snippet can access the snippet copy view and copy that source object. The copy operation reveals the contents of a snippet that the user is not authorized to view. This issue is fixed in versions 7.0.9, 7.3.4, 7.4.3, and 8.0rc2.

    Other Web app

    6.5 MEDIUM 0.3% exploit odds Fix listed
  32. CVE-2026-123820

    Wagtail is an open source content management system built on Django. Prior to 7.0.9, 7.3.4, 7.4.3, and 8.0rc2, the Wagtail Documents and Images API V2 excludes directly restricted collections but fails to exclude their descendant collections, even though descendants inherit ancestor view restrictions. An API user can therefore list the filenames and names of documents and images stored in descendant collections that should remain private. This issue is fixed in versions 7.0.9, 7.3.4, 7.4.3, and 8.0rc2.

    Other Web app

    5.3 MEDIUM 0.3% exploit odds Fix listed
  33. CVE-2026-123810

    Dynatrace MCP Server is an MCP server for Dynatrace Observability. Prior to 2.1.1, several read capabilities interpolate caller-controlled values into DQL without sufficient validation or escaping. The src/capabilities/find-monitored-entity-by-name.ts capability embeds entityNames in string searches, src/capabilities/list-problems.ts, src/capabilities/list-vulnerabilities.ts, and src/capabilities/list-exceptions.ts embed timeframe values, and src/capabilities/get-events-for-cluster.ts embeds timeframe, clusterId, and kubernetesEntityId. Crafted values can close string literals, append pipeline stages, or comment out the remaining intended query, allowing callers to bypass field selection, time-window, and result-display limits in tools marked readOnlyHint. The issue does not grant data access beyond the configured token's existing execute_dql capability, but it breaks the narrower auto-approved tool contract. This issue is fixed in version 2.1.1.

    Injection Server

    4.3 MEDIUM 0.2% exploit odds Fix listed
  34. CVE-2026-123815

    pyLoad is a free and open-source download manager written in Python. Prior to 0.5.0b3.dev102, UnTar._safe_extractall in src/pyload/plugins/extractors/UnTar.py validates archive member paths and link targets but does not call TarInfo.isdev to reject special tar member types before calling tarfile.extractall. On supported Python 3.9 through 3.13 runtimes, tarfile defaults to the fully_trusted extraction filter, so a user with ADD and STATUS permissions can cause extraction of a crafted archive containing character devices, block devices, or FIFOs. Python 3.14 defaults to a safe extraction filter. When pyLoad runs with operating-system privileges that permit device creation, including the official root-based Docker deployment, the archive can expose raw devices and compromise the host. Non-root deployments can still create FIFOs that cause hangs or IPC confusion. This issue is fixed in version 0.5.0b3.dev102.

    Disclosure Server

    9.4 CRITICAL 0.4% exploit odds Fix listed
  35. CVE-2026-123811

    Dynatrace MCP Server is an MCP server for Dynatrace Observability. Prior to 2.0.0, a server started with the --http option accepts JSON-RPC tools/call requests without authentication, session validation, or Host and Origin protections before dispatching them under the server's configured Dynatrace credentials. A network-reachable caller can invoke execute_dql to read logs, security events, user sessions, and other Grail data, or invoke create_dynatrace_notebook to create tenant-visible notebooks. The send_event path is not affected because its requestHumanApproval gate blocks this attack path, and reset_grail_budget only demonstrates dispatch without accessing Dynatrace data. This issue is fixed in version 2.0.0.

    Auth & access Server

    7.5 HIGH 0.2% exploit odds Fix listed
  36. CVE-2026-123812

    Dynatrace MCP Server is an MCP server for Dynatrace Observability. Prior to 1.8.7, the create_dynatrace_notebook registration in src/index.ts performs a persistent document-write action without calling requestHumanApproval, although the project's other write tools use that gate to obtain operator consent. A caller can create tenant-visible notebooks containing attacker-controlled Markdown or DQL content, and tenant users may later execute embedded DQL when opening the notebook. The tool also lacks a destructiveHint annotation that would signal its side effects to MCP clients. This issue is fixed in version 1.8.7.

    Auth & access Device

    3.7 LOW 0.2% exploit odds Fix listed
  37. CVE-2026-123813

    pyLoad is a free and open-source download manager written in Python. Prior to 0.5.0b3.dev102, pyLoad does not invalidate existing Flask sessions when an administrator changes a user's permissions or password through /api/set_user_permission or /api/change_password. Api.set_user_permission and Api.change_password update the database, but session authorization continues to use role and permission values copied at login by set_session in webui/app/helpers.py. A user can therefore retain revoked privileges or continue using a session established with an old password for the configured session lifetime. API-key authentication is not affected because that path rereads user data from the database. This issue is fixed in version 0.5.0b3.dev102.

    Auth & access Server

    7.5 HIGH 0.3% exploit odds Fix listed
  38. CVE-2026-123814

    pyLoad is a free and open-source download manager written in Python. Prior to 0.5.0b3.dev102, an authenticated non-admin user with the SETTINGS permission can configure WindowsPhoneNotify through /json/save_config by setting pushurl, pushid, and enabled. WindowsPhoneNotify.send in src/pyload/plugins/addons/WindowsPhoneNotify.py then uses http.client.HTTPConnection without pyLoad's pycurl destination validation. When a download completes, the global notification event causes a blind POST from the pyLoad server to the attacker-selected internal host, port, and path. The response is not returned, but the request can probe internal services, reach cloud metadata endpoints, or trigger state-changing internal HTTP endpoints. This issue is fixed in version 0.5.0b3.dev102.

    Disclosure Server

    6.4 MEDIUM 0.1% exploit odds Fix listed
  39. CVE-2026-123806

    python-tuf is a Python reference implementation of The Update Framework (TUF). Prior to 7.0.0, DelegatedRole._is_target_in_pathpattern in tuf/api/_payload.py uses fnmatch.fnmatch, which applies os.path.normcase before matching and makes delegation path authorization case-insensitive on Windows. When a TUF repository has path-based delegations that differ only by case and an attacker-controlled role is visited before the legitimate role, ngclient can accept an attacker-provided target from the wrong delegation for a requested path. Linux and macOS clients are not affected because their path normalization does not change case. This issue is fixed in 7.0.0.

    Other OS & kernel

    4.0 MEDIUM 0.1% exploit odds Fix listed
  40. CVE-2026-123807

    Signal K Server is a server application that runs on a central hub in a boat. From 2.0.0 until 2.34.0 for Signal K Server and from 1.0.0 until 1.5.3 for @signalk/resources-provider, the v2 REST route GET /signalk/v2/api/resources/:resourceType/:resourceId accepts an unsafe resourceId that can escape the intended resource directory. The route passes the attacker-controlled resourceId through src/api/resources/index.ts to FileStore.getResource() in packages/resources-provider-plugin/src/filestorage.ts without adequate path containment, allowing arbitrary JSON files reachable by the server to be read. Related PUT and DELETE resource operations can write or delete files outside the storage root. Deployments with allow_readonly enabled permit anonymous reads, while other deployments require a readonly user token. Reading security.json exposes the HS256 JWT signing secret and password hashes, allowing an attacker to forge an administrator token. Administrative access to POST /skServer/providers can then create an Execute provider that runs commands with the server process privileges and persists its configuration across restarts. This issue is fixed in Signal K Server 2.34.0 and @signalk/resources-provider 1.5.3.

    Disclosure Web app

    9.8 CRITICAL 2% exploit odds Fix listed
  41. CVE-2026-123808

    Nuxt UI is an intuitive Vue UI library powered by Reka UI and Tailwind CSS. Prior to 4.8.1, the server-rendered form element produced by UForm in src/runtime/components/Form.vue omits the method and action attributes and relies solely on the hydrated @submit.prevent handler to stop native submission. UAuthForm composes UForm and can render named credential fields, including an input with type password, inside the same markup. If a user submits before Vue hydration completes, or when the client bundle is blocked or unavailable, the browser performs its default GET submission and serializes the named fields into the current URL query string. The credentials can then appear in the address bar, browser history, reverse-proxy, CDN, or WAF request logs, and Referer headers sent with same-origin subresource requests. This issue is fixed in version 4.8.1.

    Disclosure Browser & mail

    5.3 MEDIUM 0.2% exploit odds Fix listed
  42. CVE-2026-123809

    Dynatrace MCP Server is an MCP server for Dynatrace Observability. Prior to 2.0.0, the create_workflow_for_notification tool inserts the caller-controlled teamName, problemType, and channel values into a Dynatrace workflow without neutralizing Jinja2 expressions. A caller can place template expressions such as event() in those values, and the workflow engine evaluates them when the persistent workflow runs. When the workflow has a valid Slack connector, the resulting Slack action can disclose event or tenant metadata to an attacker-selected channel; the shipped placeholder connection does not deliver messages by itself. The workflow remains in the tenant after the MCP session and credentials are gone. The isPrivate option defaults to false, making the created workflow tenant-visible unless the caller changes it. Exploitation requires either an MCP prompt-injection path or an operator approving a prompt that displays the template fragment literally. This issue is fixed in version 2.0.0.

    Injection Server

    4.2 MEDIUM 0.2% exploit odds Fix listed
  43. CVE-2026-122056

    A flaw was found in OpenSCAP. A heap-based buffer overflow can occur when processing a specially crafted Extensible Configuration Checklist Description Format (XCCDF) document using the sds-compose option. An attacker could exploit this vulnerability to corrupt memory, resulting in an application crash and a Denial of Service (DoS).

    Memory safety

    5.9 MEDIUM 0.3% exploit odds Fix listed
  44. CVE-2026-119759

    TrailBase 0.28.2 before 0.35.10 contains an insecure direct object reference in std.FileUpload columns, where the Record API accepts attacker-supplied object-store file ids on insert and update without ownership validation. An authenticated attacker who can read a victim’s row can recover its stripped file id via a $like filter and then permanently delete or orphan the victim’s file by deleting or nulling a record they own.

    Auth & access

    8.6 HIGH 0.2% exploit odds No fix listed
  45. CVE-2026-115864

    Monoceros versions before 0.30 for Perl allow a request declaring a large Content-Length to kill a worker by exhausting memory. The request body is read in a loop that asks sysread for all of the remaining declared length at once. Perl grows the read buffer to that size before any data arrives, so a length larger than available memory makes the worker exit with an uncatchable out of memory error before the application is called. A remote attacker can force a worker restart with each request of a few bytes.

    Other

    7.5 HIGH 0.6% exploit odds Fix listed
  46. CVE-2026-113971

    This issue was addressed by restricting options offered on a locked device. This issue is fixed in iOS 18.7.12 and iPadOS 18.7.12, iOS 26.5 and iPadOS 26.5. An attacker with physical access may be able to pair a device with an AirPlay receiver from the lock screen.

    Other OS & kernel

    3.5 LOW 0.2% exploit odds Fix listed
  47. CVE-2026-114031

    Incorrect permission assignment allows local privilege escalation to SYSTEM on affected releases of the Illumio VEN for Windows. Code execution as NT AUTHORITY\LOCAL SERVICE on the Windows host is required to exploit this vulnerability. The flaw exists within the VEN installer, which grants LOCAL SERVICE full access to the VEN data and installation folders, whose configuration and state files are read and trusted by the SYSTEM-privileged illumio-ven-ctl component. An attacker can leverage this vulnerability to modify the VEN configuration so that illumio-ven-ctl executes an injected command in the context of the SYSTEM account.

    Auth & access OS & kernel

    7.8 HIGH 0.1% exploit odds No fix listed
  48. CVE-2026-114032

    The OpenSSL library bundled with the Illumio VEN is affected by an uncontrolled search path vulnerability. A low-privileged, authenticated local attacker could potentially leverage this vulnerability to execute arbitrary code with elevated privileges.

    Code execution Library

    7.8 HIGH 0.1% exploit odds No fix listed
  49. CVE-2026-113963

    An authorization issue was addressed with improved state management. This issue is fixed in HomeView Software Version 27.2, iOS 27.2.1 and iPadOS 27.2.1, tvOS 27.2.1, visionOS 27.2.1, watchOS 27.2.1. An app may be able to access sensitive user data.

    Auth & access Device

    5.5 MEDIUM 0.1% exploit odds Fix listed
  50. CVE-2026-113964

    A logic issue was addressed with improved state management. This issue is fixed in HomeView Software Version 27.2, iOS 18.7.12 and iPadOS 18.7.12, iOS 26.7.3 and iPadOS 26.7.3, iOS 27.2.1 and iPadOS 27.2.1, macOS Golden Gate 27.2.1, macOS Sequoia 15.8.3, macOS Tahoe 26.7.3, tvOS 27.2.1, visionOS 27.2.1, watchOS 26.6.1, watchOS 27.2.1. An attacker in Bluetooth range may be able to bypass the rate limit on pairing code attempts and pair with a device without user interaction.

    Auth & access Device

    5.4 MEDIUM 0.3% exploit odds Fix listed

The newest published CVE records from CVE.org and NVD, refreshed every few minutes, with severity, exploit odds and fix status.