The internet is broken right now.
65 critical CVEs without a fix in the last 48 hours. Developers are crying quietly into their IDEs. 153 more already have a fix that someone still has to apply.
-
CVE-2026-125204
FFmpeg through 9.0.2 contains a heap buffer overflow in ff_libwebp_get_frame() in libwebpenc_common.c that lets attackers corrupt memory when cr_threshold is set. Attackers can supply crafted media that increases resolution mid-stream, causing memcpy() to write pixel data past the stale reference frame and crash applications.
-
CVE-2026-125196
FFmpeg through 9.0.2 contains a heap buffer overflow vulnerability in liboapve_encode() in libavcodec/liboapvenc.c that allows attackers to corrupt heap memory by supplying frames larger than the configured encoder dimensions. Attackers can craft videos whose resolution changes mid-stream so that transcoding to APV without rescaling writes attacker-controlled pixel data past the imgb plane buffers.
-
CVE-2026-123199
eBook-speaker 6.1.0 through 6.2.0 contains a stack buffer overflow vulnerability in get_attributes() in src/common.c caused by an unbounded strcpy of an element's name attribute. Attackers can craft an EPUB or DAISY book with a name attribute of 511 or more characters to corrupt the stack and crash the reader when opened.
-
CVE-2026-93670
The Synopsys DesignWare I3C controller driver (drivers/i3c/i3c_dw.c) tracks free Device Address Table (DAT) slots with get_free_pos(), which signals an exhausted table by returning -1. The helper declared its return type as uint8_t, so the sentinel was truncated to 255, and both consumers (dw_i3c_attach_device() and dw_i3c_i2c_attach_device()) also stored it in a uint8_t before testing if (pos < 0) — a comparison that is never true for an unsigned value. The exhaustion guard was therefore inert in every build. With the guard defeated, an attach attempted on a full DAT proceeds with pos == 255: it writes data->dw_i3c_i2c_priv_data[255].id, more than 200 bytes past the end of the 32-element dw_i3c_i2c_priv_data array and past the end of struct dw_i3c_data; it stores the resulting out-of-bounds address in desc->controller_priv, from which later detach, reattach and IBI operations read the slot index back; and it performs a 32-bit MMIO store 1020 bytes beyond the hardware DAT base, outside the table the controller reports. The attach still returns success, so the upper layer believes the device is bound. The DAT is smaller than the software array (its depth is read from DEVICE_ADDR_TABLE_POINTER, commonly 8 or 16 entries), so exhaustion is reachable. In a dual-role build (CONFIG_I3C_DUAL_ROLE with CONFIG_I3C_USE_IBI) acting as a secondary controller, the attach count is supplied by the bus's active controller: i3c_sec_handoffed() attaches one device per entry of the received DEFTGTS list, so a hostile or malfunctioning controller on the local I3C bus that declares more targets than the local DAT depth triggers the condition once per surplus entry. Exhaustion can also be reached from a devicetree that declares more devices than the table holds, or by repeated i3c attach shell commands. Impact is bounded: the out-of-bounds store is a single byte of fixed value at a fixed displacement, with no attacker-chosen data or offset, so the realistic outcomes are corruption of neighbouring static driver state, writes into reserved controller registers, and inconsistent driver bookkeeping (a crash or a wedged I3C bus) rather than a controllable memory-write primitive. The attack vector is the local chip-to-chip I3C bus; there is no network-reachable path.
-
CVE-2026-127100
A flaw was found in SSSD's PAM responder on builds with the JSON PAM extension enabled. A local attacker can submit a crafted authentication request containing a non-string OAuth2 code value. Missing JSON type validation leads to a NULL pointer dereference, terminating the PAM responder and disrupting SSSD-backed authentication.
-
CVE-2026-32667
Out-of-bounds write for the Intel(R) VTune(TM) Profiler software within Ring 1: Device Drivers may allow an escalation of privilege. Unprivileged software adversary with an authenticated user combined with a low complexity attack may enable escalation of privilege. This result may potentially occur via local access when attack requirements are not present without special internal knowledge and requires no user interaction. The potential vulnerability may impact the confidentiality (high), integrity (high) and availability (high) of the vulnerable system, resulting in subsequent system confidentiality (none), integrity (none) and availability (none) impacts.
-
CVE-2026-125241
FFmpeg through 9.0.2 contains a heap buffer overflow vulnerability in libavcodec/flicvideo.c that allows attackers to write out of bounds because check_pixel_ptr() checks only one bound with negative-stride frames. Attackers can supply a crafted FLIC file whose RLE and delta chunks write pixel data past frame storage, corrupting heap memory to crash applications or possibly execute code.
-
CVE-2026-125233
FFmpeg through 9.0.2 contains a heap buffer overflow vulnerability in dvbsub_encode() in libavcodec/dvbsubenc.c that undercounts each CLUT segment header by two bytes. Attackers can supply a file with many bitmap subtitle rectangles that, when transcoded to the dvbsub encoder, writes past the output buffer and corrupts heap memory.
-
CVE-2026-125209
FFmpeg before 8.1.3 and 9.x before 9.0.2 contains a heap buffer overflow vulnerability in the DVD subtitle encoder that underestimates RLE output size for odd-width rectangles. Attackers can supply files with crafted bitmap subtitles that, when transcoded to the dvdsub encoder, write past the output buffer and corrupt heap memory.
-
CVE-2026-125203
FFmpeg before 8.1.3 and 9.x before 9.0.2 contains a heap buffer overflow vulnerability in the Cook multichannel decoder in libavcodec/cook.c that fails to verify subpacket channel masks are disjoint. Attackers can supply a crafted RealMedia Cook stream with overlapping subpacket masks, causing cook_decode_frame() to write decoded samples beyond allocated output planes in any application decoding the file.
-
CVE-2026-113780
ESF-IDF is the Espressif Internet of Things (IOT) Development Framework. From 5.5.5 until 5.5.6, 6.0.4, and 6.1.1, the esp_tee trap entry in esp_tee_vectors_plic.S and esp_tee_vectors_clic.S saves a caller-controlled register frame through the REE-supplied stack pointer before switching to the trusted TEE stack or validating the pointer. A local unprivileged REE application can direct this machine-mode write into TEE-exclusive memory on ESP32-C5, ESP32-C6, ESP32-C61, and ESP32-H2 devices, breaking the REE/TEE isolation boundary and potentially corrupting TEE control-flow data for code execution, denial of service, or system crashes. Both secure-service ecall and interrupt entry paths contain the vulnerable ordering. This issue is fixed in version 5.5.6 and is scheduled to be fixed in versions 6.0.4 and 6.1.1.
-
CVE-2026-112922
A flaw was found in Ghostscript. A remote attacker could exploit this vulnerability by enticing a user to process a specially crafted PDF or PostScript document containing malicious mesh shading data. Due to improper bounds validation during the processing of indexed color values, Ghostscript triggers an out-of-bounds memory read. Successful exploitation can lead to a Denial of Service (DoS) through an application crash or result in limited information disclosure.
-
CVE-2026-94024
The Module Manager's object lookup validated one byte of a memory-protected module's search name and then walked as much of it as the module declared, in privileged mode, with the kernel's preemption count raised. Both lookup dispatchers are affected: the recommended `txm_module_object_pointer_get_extended` is handed a length and does not use it to bound the buffer, and the deprecated `txm_module_object_pointer_get` supplies the largest value a `UINT` can hold, so it removes the bound rather than lacking one. A module can make the walk unbounded without knowing anything about the memory it reaches, and where the walk reaches memory a privileged read cannot take, the resulting fault leaves the kernel unable to preempt for the rest of the system's life. Vulnerable functions: `_txm_module_manager_txm_module_object_pointer_get_dispatch`, `_txm_module_manager_txm_module_object_pointer_get_extended_dispatch` (both in `common_modules/module_manager/inc/txm_module_manager_dispatch.h`) and `_txm_module_manager_object_name_compare` (`common_modules/module_manager/src/txm_module_manager_util.c`).
-
CVE-2026-95111
A memory-protected ThreadX module's queue request reaches the Module Manager's dispatch layer, which decides how much of the module's buffer the privileged queue copy may touch and hands that extent to the module port's range check. A queue's `tx_queue_message_size` is a count of `ULONG`s and the copy moves that many words, so the extent is `message_size * sizeof(ULONG)` bytes. The front-send dispatcher passed the word count itself. `tx_queue_send` and `tx_queue_receive`, either side of it in the same header, have always multiplied. A source buffer ending `message_size` bytes into memory the module may read was therefore accepted, and `_txe_queue_front_send` then read `message_size * sizeof(ULONG)` bytes from it in privileged mode — `3 * message_size` bytes past the end of what the module is allowed to read. The words are written into the queue's message area, so a module that can receive from that queue reads them back.
-
CVE-2026-120758
HDF5 through 2.2.0 contains a NULL pointer dereference vulnerability in H5HL_unprotect() in src/H5HL.c. H5HL_protect() pins a single metadata cache entry for a local heap, either the heap prefix or the separate data block, and H5HL_unprotect() unpins it on the final unprotect. When the metadata cache destroys a corrupted local heap entry, H5HL__prfx_dest() unlinks it from the heap by setting heap->prfx or heap->dblk to NULL. A crafted HDF5 file can therefore reach H5HL_unprotect() with no cached object to unpin, and the function passes the NULL pointer to H5AC_unpin_entry(). This causes an assertion failure in debug builds and a NULL pointer dereference and crash in release builds. An attacker who can get a user or application to open and traverse a malicious file can cause a denial of service.
-
CVE-2026-120755
HDF5 through 2.2.0 contains an out-of-bounds read vulnerability in H5Z__filter_szip() in src/H5Zszip.c. When decoding a szip-compressed chunk, the function subtracts the 4-byte uncompressed-length header from the chunk length without first verifying that the chunk is at least 4 bytes long. A truncated or corrupted chunk shorter than the header causes the length calculation to wrap around to a very large value, and the decompressor reads far past the end of the chunk buffer. An attacker who can get a user or application to read a dataset from a crafted HDF5 file can cause a denial of service. Only builds with szip support (libaec or the szip library) are affected.
-
CVE-2026-120438
QuickJS-NG before 0.16.0 contains an integer overflow in js_proxy_get_own_property_names() in quickjs.c that causes a heap buffer overflow on 32-bit hosts. Attackers running JavaScript can return a large array-like object from a Proxy ownKeys trap to wrap the allocation size, causing crashes or potential code execution.
-
CVE-2026-107894
V-SFT provided by FUJI ELECTRIC CO., LTD. contains a stack-based buffer overflow vulnerability. Opening a crafted V7 file may lead to information exposure or arbitrary code execution on the affected product.
-
CVE-2026-21153
Stack-based buffer overflow in libsecuibc.so prior to SMR Nov-2026 Release 1 allows local attackers to execute arbitrary code.
-
CVE-2026-63418
Monitoring Plugins is a suite of plugins for monitoring applications. From 3.0.0-rc1 until 3.0.1, check_icmp stores the number of hosts supplied through -H or --Host on the command line, or through --extra-opts files containing Host= entries, in an unsigned short. More than 65,535 hosts wrap the counter, causing process_arguments() to allocate an undersized hosts array and then write subsequent entries out of bounds before main() calls setuid(getuid()). On installations where check_icmp is setuid root, a local low-privileged attacker can corrupt heap memory while the process still has root privileges; reliable privilege escalation has not been demonstrated. Installations that use cap_net_raw file capabilities instead of setuid are not exposed to the privilege-escalation consequence. This issue is fixed in version 3.0.1.
-
CVE-2026-65108
NVIDIA Slurm contains an attacker-controlled heap-based buffer overflow in the slurmstepd process while the slurmstepd is running with root permissions.
-
CVE-2026-84599
The issue was addressed with improved memory handling. This issue is fixed in iOS 26.7.2 and iPadOS 26.7.2, iOS 27 and iPadOS 27, macOS Golden Gate 27, macOS Sequoia 15.8, macOS Tahoe 26.7, tvOS 27, visionOS 27, watchOS 27. An app may be able to cause unexpected system termination or corrupt kernel memory.
-
CVE-2026-116497
A NULL pointer dereference vulnerability was found in gssproxy. When a GSSX_INIT_SEC_CONTEXT request is received with the target_name field absent, the daemon dereferences the resulting NULL pointer and crashes. The request requires no authentication, allowing any local user to crash the gssproxy daemon via its Unix socket. The daemon does not automatically restart, resulting in a persistent denial of service that also affects dependent services such as Kerberized NFS and IdM.
-
CVE-2026-120866
Memory overflow / improper handling in BGP dynamic routing can lead to unpredictable or erroneous behavior or Denial of Service in Citrix NetScaler ADC and Citrix NetScaler Gateway.
-
CVE-2026-119338
Tesseract is an open source OCR engine. From 5.4.0 until 5.5.4, AppendLinePolygon in src/api/pagerenderer.cpp destroys the caller-owned right-to-left Pta accumulator while receiving it by value, leaving line_top_rtl_pts and line_bottom_rtl_pts dangling. On an upright Hebrew or Arabic page at the default PAGE XML level, POLYGONFLAG with a false skewed_flag enters the vulnerable calls, while LEVELFLAG greater than zero provides an alternative gate. The calls repeatedly read and free the stale object, causing a double free and use after free and allowing corrupted coordinates to reach RecalcPolygonline. This issue is fixed in version 5.5.4.
-
CVE-2026-99109
In the Linux kernel, the following vulnerability has been resolved: smb: client: reject SetEA requests that do not fit the request buffer CIFSSMBSetEA() copies the caller's extended attribute value into the SMB request buffer without checking that it fits. The requirement is stated in the source but was never implemented: /*BB add length check to see if it would fit in negotiated SMB buffer size BB */ /* if (ea_value_len > buffer_size - 512 (enough for header)) */ if (ea_value_len) memcpy(parm_data->list.name + name_len + 1, ea_value, ea_value_len); The only bound applied on the way in is in cifs_xattr_set(): #define MAX_EA_VALUE_SIZE CIFSMaxBufSize ... if (size > MAX_EA_VALUE_SIZE) CIFSMaxBufSize is the full payload capacity of the buffer, so a value of exactly that size leaves no room for the SMB header, the TRANS2 parameter block, the fealist header and the EA name that are written ahead of it in the same object. SendReceive() already enforces the correct limit on this very length: if (in_len > CIFSMaxBufSize + MAX_CIFS_HDR_SIZE) but it is called after the copy has taken place. An unprivileged setxattr(2) on an SMB1 mount with a 250-byte name and a 16384-byte value writes 16384 bytes starting 345 bytes into a 16588-byte cifs_request object, ending 141 bytes past it: BUG: KASAN: slab-out-of-bounds in CIFSSMBSetEA+0xabc/0xde0 Write of size 16384 at addr ffff888003aa0159 by task init/68 __asan_memcpy+0x3c/0x60 CIFSSMBSetEA+0xabc/0xde0 cifs_xattr_set+0xd3a/0xff0 __vfs_setxattr+0x13e/0x1a0 The buggy address is located 345 bytes inside of allocated 16588-byte region Apply SendReceive()'s limit to the assembled request before the copy rather than after it, and widen the byte counters so the sum cannot wrap before it is tested. byte_count is also tested against U16_MAX, because it is stored in the 16-bit pSMB->ByteCount. That becomes reachable when CIFSMaxBufSize is raised at module load, where it may be set as high as 1024*127: with a 5-byte EA name and a 65521-byte value, count is exactly U16_MAX while byte_count is 65556, and cpu_to_le16() would truncate it to 20 and transmit a frame whose ByteCount does not match its length. Testing byte_count covers count as well, since byte_count is the larger of the two and count's only 16-bit consumer is written after this point. check_add_overflow() is evaluated first so that total_len is assigned before it is reported.
-
CVE-2026-113945
An out-of-bounds read was addressed with improved bounds checking. This issue is fixed in macOS Golden Gate 27.2. An app may be able to cause unexpected system termination.
-
CVE-2026-16662
boot_save_boot_status() in boot/bootutil/src/boot_record.c reads the image's EXPECTED_HASH_TLV into a stack buffer image_hash[IMAGE_HASH_SIZE] that is never zero-initialised. The TLV handler validates only the upper bound (len > sizeof(image_hash)) and then reads exactly len bytes, so a TLV shorter than IMAGE_HASH_SIZE leaves the tail of image_hash holding stale stack content. The subsequent copy into the boot record, memcpy(buf + offset, image_hash, sizeof(image_hash)), uses the fixed buffer size rather than len, so up to IMAGE_HASH_SIZE - 1 bytes of uninitialised bootloader stack are written into the measured-boot record. The defect is reachable only under MCUBOOT_MEASURED_BOOT, and only when nothing has checked the length of the hash TLV before boot_save_boot_status() reads it. bootutil_img_validate() rejects a hash TLV whose length is not exactly sizeof(hash) (boot/bootutil/src/image_validate.c:345), and primary-slot validation is on by default (Zephyr's CONFIG_BOOT_VALIDATE_SLOT0 defaults to y), so in a default build the malformed image never reaches this code. Two kinds of build let a short TLV through: Builds that use pure signatures (MCUBOOT_SIGN_PURE, Zephyr CONFIG_BOOT_SIGNATURE_TYPE_PURE). The signature covers the image itself, so the hash-TLV check in bootutil_img_validate() is compiled out, and a pure-signed image normally carries no hash TLV. A hash TLV is still allowed in the unprotected TLV area, which the signature does not cover, so a short one can be added to a correctly signed image without invalidating it. This applies with every upgrade strategy, including direct-XIP and RAM-load, and the modified image can be delivered through any path that accepts a signed image. Builds with primary-slot validation turned off: the swap and overwrite loader built without MCUBOOT_VALIDATE_PRIMARY_SLOT, which checks only the image magic before calling boot_add_shared_data(), and the Zephyr single-slot loader (boot/zephyr/single_loader.c) built without it, including the MCUBOOT_VALIDATE_PRIMARY_SLOT_ONCE variant, which skips validation once the trailer records an earlier successful one. Placing the crafted image requires physical or flash-write access to the device; in a single-slot build with serial recovery enabled, the serial-recovery upload writes the primary slot directly and can place it as well. Outside pure-signature builds, the direct-XIP and RAM-load strategies in boot/bootutil/src/loader.c validate the selected slot unconditionally and are not affected. It is also reachable through a time-of-check/time-of-use change to tamperable external flash between validation and the boot-record read. In these configurations the boot record's measurement is already not trustworthy: the hash copied into the record is the image's own EXPECTED_HASH_TLV, which nothing has verified and which the attacker chooses. The defect therefore does not weaken attestation integrity any further. What it adds is a path for bootloader memory to leave the device: up to 63 bytes (IMAGE_HASH_SIZE - 1) of uninitialised bootloader stack are embedded in the measurement field that a downstream attestation service sends to a remote verifier. In the configurations analysed the residue is bootloader internal state (TLV offsets and lengths, flash addresses, loop state), not key material, but what occupies that stack region depends on configuration and on which code ran before boot_save_boot_status(). It is reported as a low-severity, defense-in-depth disclosure with no integrity or availability impact. This is distinct from the already-fixed record_len integer underflow in the same routine (GHSA-8hrv-4cp5-4rg3). The fix rejects any EXPECTED_HASH_TLV whose length is not exactly IMAGE_HASH_SIZE.
-
CVE-2026-125267
FFmpeg before 8.1.3 and 9.x before 9.0.2 contains a heap out-of-bounds write vulnerability in the APAC decoder in libavcodec/apac.c caused by an unreset bitstream index. Attackers can supply a crafted APAC file whose growing bitstream_index triggers an integer overflow, bypassing the reallocation check so memcpy() corrupts heap memory or crashes the application.
-
CVE-2026-120252
imagecodecs 2023.1.23 through 2026.10.10 contains a stack-based buffer overflow vulnerability in the lossless JPEG Define Huffman Table parser in decode_jpegsof3(). Attackers can supply a crafted lossless JPEG whose DHT symbol counts wrap an 8-bit sum, overflowing the 32-entry HufVal and HufSz stack arrays and crashing the calling process.
-
CVE-2026-0455
The Versal™ AI Edge Series Gen 2 Peripheral Protection Unit (XPPU) for the inter-processor interrupt (IPI) buffers does not block unauthorized access as expected. This allows processors in the SoC with any IPI channel access the ability to both read and write to the entire IPI buffer register space as well as the IPI no-buffer register control space.
-
CVE-2026-116731
Furnace through 0.6.8.3 contains an out-of-bounds write vulnerability in DivEngine::loadFTM() that allows attackers to corrupt heap memory via an unchecked subsong index in the GROOVES block. Attackers can craft a FamiTracker module with a subsong count of 2 or more to write a groove pattern past the subsong vector, crashing Furnace when opened.
-
CVE-2026-116725
gbsplay through 0.0.103 contains an out-of-bounds read vulnerability in mapper_map() in mapper.c that fails to reject ROM bank indexes beyond the ROM size, causing an underflowed bank length. Attackers can supply crafted .gbr, .gbs or .gb files with malicious bank indexes to make bank_get() read far past the heap ROM buffer and crash the player.
-
CVE-2026-116728
libfvde through 20260923 contains a heap out-of-bounds read vulnerability in libfvde_encrypted_metadata_read_type_0x0024() because xml_plist_data_size is not checked against the source metadata block. Attackers can supply a crafted FileVault 2 Core Storage volume that makes libfvde_volume_open() copy past the heap block buffer, crashing the application.
-
CVE-2026-116730
Furnace through 0.6.8.3 contains an out-of-bounds heap read vulnerability in DivEngine::loadFTM() in src/engine/fileOps/ftm.cpp, triggered by an unbounded DPCM note count during FamiTracker module import. Attackers can craft a .ftm, .dnm or 0CC/E-FamiTracker module with a 2A03 instrument note count above 120 to read past the noteMap array and crash Furnace when opened.
-
CVE-2026-129606
Buffer Copy without Checking Size of Input ('Classic Buffer Overflow') vulnerability in Samsung Open Source mTower allows Overflow Buffers.
-
CVE-2026-115501
IBM i 7.3 7.4, 7.5, and 7.6 is vulnerable to a stack buffer overflow, caused by improper bounds checking in the QFileSvr.400 file system. By sending a command in the QFileSvr.400 file system, a local authenticated attacker could cause undefined behavior.
-
CVE-2026-40424
Out-of-bounds read for some Hyperscan before version v5.4.2.1 within Ring 3: User Applications may allow a denial of service. Unprivileged software adversary with an unauthenticated user combined with a low complexity attack may enable denial of service. This result may potentially occur via local access when attack requirements are present without special internal knowledge and requires no user interaction. The potential vulnerability may impact the confidentiality (low), integrity (none) and availability (high) of the vulnerable system, resulting in subsequent system confidentiality (none), integrity (none) and availability (none) impacts.
-
CVE-2026-125197
FFmpeg through 9.0.2 contains an integer underflow in decode_frame() in libavcodec/lcldec.c that allows attackers to trigger a heap out-of-bounds read via short ZLIB packets with FLAG_MULTITHREAD. Attackers can supply a crafted AVI or MKV file with an LCL video stream whose packet is shorter than 8 bytes, bypassing length clamping to crash the decoding application.
-
CVE-2026-112109
A local privilege escalation vulnerability in Thales SafeNet Authentication Client (SAC) SACSrv can allow an authenticated, non-administrator local user to trigger a heap-based buffer overflow through the service interface. Successful exploitation may allow arbitrary code execution with Local System privileges.This issue affects SAC releases up to and including 10.9 R3 GA (build 10.9.7449).
-
CVE-2026-40652
An out-of-bounds read condition in the AMD Compressonator SDK could allow an attacker with local access to supply a specially crafted Truevision Advanced Raster Graphics Adapter (TGA) file to a user or an automated pipeline, potentially resulting in disclosure of sensitive heap memory contents or denial of service.
-
CVE-2026-40659
An integer overflow condition in the AMD Compressonator SDK could allow an attacker with local access to supply a specially crafted DirectDraw Surface (DDS) file, potentially resulting in denial of service (DoS).
-
CVE-2026-40660
A heap-based out-of-bounds (OOB) read vulnerability in the AMD Compressonator SDK could allow an attacker with local access to read out-of-bounds heap memory, potentially resulting in information disclosure or application crash.
-
CVE-2026-40661
An out-of-bounds write (OOW) condition in the AMD Compressonator SDK could allow an attacker with local access to supply a specially crafted DirectDraw Surface (DDS) file to a user or an automated pipeline, potentially resulting in heap corruption or denial of service (DoS).
-
CVE-2026-20550
In gnss, there is a possible out of bounds write due to a missing bounds check. This could lead to local denial of service with no additional execution privileges needed. User interaction is needed for exploitation. Patch ID: ALPS11309816; Issue ID: MSV-10731.
-
CVE-2026-114690
A flaw was found in GCC. An out-of-bounds memory access vulnerability exists in the libstdc++ parallel search implementation due to an undersized internal buffer. A local attacker can exploit this flaw by providing crafted search data to an application compiled with parallel mode, causing the process to crash and leading to a Denial of Service (DoS).
-
CVE-2026-99142
In the Linux kernel, the following vulnerability has been resolved: tracing: Check operator stack capacity before pushing predicate_parse() uses the first operator stack entry before processing any opening parentheses. A push then increments top before writing. The existing check still permits that write when top already points to the last allocated entry, which can access beyond the operator stack. Reject the push when top reaches the last allocated entry.
-
CVE-2026-117608
Assimp through 6.0.5 contains an integer overflow vulnerability in MD2Importer::ValidateHeader() that allows attackers to bypass frame table bounds checks using crafted header fields. Attackers can supply a malicious .md2 file whose numFrames and numVertices values wrap the 32-bit size check, causing heap out-of-bounds reads that crash applications or expose adjacent heap memory.
-
CVE-2026-116816
llama.cpp b5899 before b10261 contains a heap out-of-bounds write vulnerability in llm_tokenizer_plamo2::build that allows attackers to corrupt heap memory via crafted GGUF models. Attackers can supply BYTE-typed token text like <0x-1> whose negative parsed value writes token ids before the bytes_ buffer during model loading, causing crashes or potential code execution.
-
CVE-2026-91189
The GUIX binary resource loader never checks that a font page's glyph range is well formed. `gx_page_header_first_glyph` and `gx_page_header_last_glyph` are read verbatim from the file and only the upper bound of `last_glyph` is validated, so a page declaring `first_glyph > last_glyph` makes `last - first + 1` underflow to a very large unsigned count. That count is multiplied by a glyph record size to reserve the load buffer, and the multiply is unchecked, so it wraps to a small value and the loader allocates a buffer far smaller than the data it is about to write. The load pass then recomputes the count in a narrower type and writes attacker-controlled glyph records into the undersized buffer.