Suse
Monthly
A race condition in the btrfs filesystem's relocation code (merge_reloc_roots() in fs/btrfs/relocation.c) allows a local user with an active balance or device add/remove/replace operation in progress to observe an inconsistent root->reloc_root pointer because the code clears that pointer without the memory barrier used elsewhere in the same code path. The affected code also clears the BTRFS_ROOT_DEAD_RELOC_TREE flag even when the root no longer points at the reloc root being cleaned up, so cleanup can target the wrong reloc root. The upstream fix routes cleanup through clear_reloc_root(), which issues the barrier after nulling ->reloc_root and before clearing the flag. Our independent assessment rates this far below the vendor-published CVSS 7.8: it is a local, high-complexity race that requires btrfs to be in active use and block-group relocation to be underway, is only meaningfully exploitable on weakly-ordered CPU architectures, and yields at most low integrity and availability impact with no confidentiality loss; there is no remote vector and no public exploit identified at time of analysis, and no KEV listing.
The Linux kernel's iwlwifi PCIe transport driver can leave dangling RX pointers (rx_pool, global_table, rxq and alloc_page) on the live transport object after iwl_pcie_rx_free() runs while unwinding a failed iwl_pcie_tx_init(), so a subsequent teardown or driver re-init/retry can touch already-freed RX state. Only systems with Intel WiFi hardware running kernels before 6.12.111, 6.18.53 or 7.2 are affected, and the trigger is an internal nic-init failure path rather than a directly attacker-controlled input, meaning exploitation is local with high complexity and existing low privileges (AV:L/AC:H/PR:L per the assessed vector). There is no public exploit identified at time of analysis and no KEV confirmation of active exploitation, and while the vendor/NVD score is 7.0 (High), the independent assessment rates this at roughly 3.6 (Low) because only integrity and availability of the device's own error/cleanup path are at stake.
Out-of-bounds reads and subsequent kernel instability in the Linux kernel's iwlwifi MVM driver (Intel WiFi) can be triggered when the WiFi firmware returns a malformed TX_CMD response whose frame_count does not match the actual response buffer layout. The affected code parses status entries by walking frame_count and then reads a trailing SCD SSN field without first validating that the buffer is long enough, so a malicious, compromised, or substituted device firmware can push the parser past the end of the payload, yielding limited information disclosure and, more likely, a kernel fault that takes down the driver or system. Systems running any Linux kernel built from the affected ranges (fixed in 6.12.111, 6.18.53, and 7.2) with an Intel WiFi adapter are exposed; there is no public exploit code and no confirmed active exploitation (CISA KEV) at time of analysis, and the practical attack surface is a firmware-to-host trust boundary rather than over-the-air radio traffic. Note that the supplied CVSS of 8.8 (AV:A/AC:L) overstates real-world risk; an independent assessment scores this AV:L/AC:H with confidentiality Low and availability High.
Out-of-bounds tid_data array access in the Linux kernel's iwlwifi mvm driver (CVE-2026-93790) is a theoretical flaw that writes the lq_color value into an unrelated tid_data slot, but it is not practically triggerable because multi-TID blockack is not in use, meaning the TFD loop counter 'i' is always 0 and the write remains in-bounds on unpatched systems. The affected component is the Intel iwlwifi mvm Wi-Fi driver in Linux kernel versions prior to 6.18.53, 6.12.111, and 7.2. An attacker within Wi-Fi radio range (adjacent, unauthenticated per the assessed CVSS vector) would need a contrived multi-TID blockack scenario to cause minor rate-control state corruption; no public exploit code exists and there is no evidence of active exploitation, so real-world risk is low.
An out-of-bounds slab read in the Linux kernel's SMB/CIFS client allows an attacker-controlled SMB1 server to make a victim's kernel read past the end of a kmalloc'd response buffer when that victim lists a directory on a mounted share. cifs_filldir() trusts the server-supplied de.namelen field from a TRANS2_FIND_FIRST/FIND_NEXT response without verifying that the name stays inside the response buffer, so an oversized FileNameLength drives memchr() (via filldir64()/verify_dirent_name()) beyond the allocation and produces a KASAN slab-out-of-bounds report; CVSS 3.1 is 8.1 (AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:N/A:H), with high confidentiality and availability impact from the out-of-bounds read. Exploitation requires the victim to mount a share from a hostile server using the deprecated SMB1 dialect and then browse it, so user interaction is required; no public exploit identified at time of analysis.
Permission widening in the Linux kernel's ksmbd in-kernel SMB3 server allows an authenticated SMB session holder to create files and directories whose inherited POSIX ACLs are recomputed incorrectly, discarding restrictive ACL_MASK entries that the parent directory's default ACL was intended to enforce, so newly created objects can end up more permissive than the parent's policy allows. The bug affects the ksmbd code shipped in Linux kernels before 6.18.53 and before 6.12.111, and in practice only matters where an SMB share exports a parent directory with a restrictive default POSIX ACL; userspace Samba deployments are unaffected. Vendor-released patches exist (6.18.53, 6.12.111 and 7.2) and no public exploit code has been identified at time of analysis.
A race condition in the Linux kernel's vhost-scsi backend allows a local, authenticated attacker with access to the vhost-scsi device to write command responses to an unrelated userspace memory region by replacing the memory table (VHOST_SET_MEM_TABLE) while SCSI commands are in flight. This can lead to host memory corruption and potential privilege escalation, though the attack requires a non-default vhost-scsi offload configuration and a narrow race window that makes reliable exploitation difficult. No public exploit has been identified at time of analysis; patches are available in Linux 6.12.111, 6.18.53, and 7.2.
A use-after-free in the Linux kernel's netfilter nfnetlink_log (NFLOG) subsystem lets a local attacker who can create and tear down network namespaces race packet logging against per-net state teardown, so that ops_free_list() frees the NFLOG pernet data while nf_log_packet() may still dereference the stale logger pointer under rcu_read_lock(). Impact is confined to local hosts running an affected kernel where the NFLOG logging path is in use and unprivileged network-namespace creation (or CAP_NET_ADMIN) is available; the flaw is a timing race, not a remotely reachable bug, and the assessed vector is AV:L/AC:H/PR:L. No public exploit code has been identified at time of analysis and the issue is not confirmed actively exploited (CISA KEV); vendor patches are available in Linux 6.12.111, 6.18.53, and 7.2.
Oversized SMBus block transfers in the Linux kernel I2C subsystem can reach tracepoints and native driver callbacks without ever being validated against I2C_SMBUS_BLOCK_MAX, allowing local memory corruption and out-of-bounds reads. The flaw is fixed by adding an early length check in __i2c_smbus_xfer(), which previously deferred validation to i2c_smbus_xfer_emulated() - too late for the smbus_write tracepoint and skipped entirely by adapters that supply a native .smbus_xfer implementation. The independently assessed vector (AV:L/AC:H/PR:L/UI:N/S:U/C:H/I:N/A:H) reflects that exploitation requires local access to an /dev/i2c-* interface plus a non-default configuration - SMBus tracing enabled (Bug 1) or the i2c-stub testing module loaded (Bug 2) - and default production kernels without tracing, i2c-stub, or unprivileged i2cdev access are not readily exploitable; no public exploit code was identified at time of analysis.
Local authenticated users on systems running Linux kernel 6.17 or newer with GPU drivers that exercise drm/pagemap device-memory migration can, on the rare migration-error path, cause DMA mappings to remain after pages are released because drm_pagemap_migrate_unmap_pages() is called after the pages array has been cleared. This can lead to limited integrity and availability impacts within the affected subsystem. The input CVSS of 8.8 substantially overstates the real-world risk; independent assessment rates it 5.3 (AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:L). No public exploit identified at time of analysis.
Out-of-bounds reads in the Linux kernel's Greybus audio topology parser allow a malicious or compromised Greybus module to leak kernel memory and crash the system. The module declares a small topology blob size but supplies large le32 section sizes (size_dais, size_controls, size_widgets) that are added to build the control, widget and route section offsets without ever being checked against the fetched blob, so gbaudio_tplg_parse_data() walks past the allocation. Exploitation requires the kernel to be built with Greybus audio support and a rogue module attached to the UniPro/Greybus interconnect - a rare hardware configuration - and no public exploit code has been identified at time of analysis. Note that while the supplied NVD-style vector rates this 8.8 with C:H/I:H/A:H under AV:A, our assessment is that the weakness is fundamentally a read-side bug requiring physical/hardware access (AV:P, C:H/I:N/A:H), and the 8.8 score overstates both the access path and the integrity impact.
Local privilege escalation-class memory-safety flaw in the Linux kernel's RDMA/bnxt_re driver allows an authenticated local user with access to the RDMA userspace-verbs interface (/dev/infiniband/uverbsN) to trigger use-after-free or double-free conditions by forcing udata copy failures after resource teardown, causing destroy callbacks to run again against freed driver state; a create-AH path can also leak a hardware object. The CVSS 3.1 vector is AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H (7.8), and exploitation requires a Broadcom NetXtreme-E adapter running bnxt_re in RoCE/RDMA mode, which is not a default configuration on general-purpose hosts. No public exploit was identified at time of analysis.
Out-of-bounds memory access in the Linux kernel PCI/pwrctrl driver for the Toshiba TC9563 PCIe switch, caused by the driver parsing both physical functions of the DSP3-integrated Ethernet MAC Endpoint when only one entry exists in the tc9563_pwrctrl->cfg array. Affected systems are those running Linux 6.19 or kernels built from commit 4c9c7be47310c1dbd7b6d37d45986123f5b133b4 up to the fixes in 7.2.6 / 7.3-rc1, but only when the specific TC9563 hardware is physically present, since the buggy code runs during driver probe; there is no network path or attacker-supplied runtime input. No public exploit has been identified at time of analysis (no CISA KEV listing, no POC, no EPSS signal provided), and while the published vector is 7.7, independent assessment rates real-world risk LOW with a modified vector of AV:L/AC:H/PR:H/UI:N/C:N/I:L/A:H, reflecting the hardware prerequisite as the dominant limiting factor.
A use-after-free in the Linux kernel's MD RAID5 Partial Parity Log (PPL) implementation lets a local attacker holding low privileges on a host that runs a PPL-enabled RAID5 array corrupt kernel memory, with the assessed impact being full loss of confidentiality, integrity and availability (CVSS 7.8, AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H) and a plausible path to kernel code execution or privilege escalation to root. The bug lives in ppl_do_flush(), whose loop keeps iterating over io->pending_flushes after ppl_io_unit_finished() has already freed the io unit; the fix is a single break statement that terminates the loop once the unit is released. Vendor patches are available across the affected stable branches, and no public exploit code was identified at time of analysis; exploitation is gated on reaching the PPL flush path at all (RAID5 with PPL journaling) and on winning the use-after-free window, so practical priority is moderate rather than urgent.
A use-after-free in the Linux kernel's VXLAN multicast database (MDB) handling allows a local attacker who holds CAP_NET_ADMIN (including within a user namespace) to corrupt kernel memory by manipulating VXLAN MDB entries and then triggering a flush. The bug lives in vxlan_mdb_flush(), which iterates the MDB hash list with hlist_for_each_entry_safe() under the false assumption that only the current entry can be removed; flushing the remotes of a (*, G) entry actually cascades into removal of the associated (S, G) entries, one of which may be the saved next-pointer, producing a double hlist_del() that writes LIST_POISON1 into LIST_POISON2 and yields attacker-influenced memory corruption (KASAN reported wild-memory-access). Per the assessed vector (CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H, 7.8) the impact is high for confidentiality, integrity and availability, but exploitation requires local privileges, CAP_NET_ADMIN, and a non-default VXLAN MDB configuration; there is no remote or unauthenticated path. No public exploit code was identified at time of analysis and the issue is not confirmed actively exploited (not in CISA KEV); vendor-released patches exist for the affected stable branches.
Page-table corruption can be triggered on LoongArch Linux kernels when the architecture fails to define DIRECT_MAP_PHYSMEM_END, causing get_free_mem_region() and mhp_get_pluggable_range() to fall back to the generic mm.h constant (1ULL << MAX_PHYSMEM_BITS) - 1 instead of the CPU's actual vmemmap-covered range. On LoongArch CPUs whose probed physical address bits (cpu_pabits from CPUCFG1) are fewer than MAX_PHYSMEM_BITS - notably the Loongson-2K series - a ZONE_DEVICE region can be allocated outside the vmemmap window, and vmemmap_populate() then wraps the memmap range around into low memory, silently corrupting page tables. The bug is local (CVSS vector PR:L), requires specific hardware plus CONFIG_SPARSEMEM_VMEMMAP and the ZONE_DEVICE/device-memory or amdkfd code path, and our independent assessment rates it AC:H with only C:L (versus the auto-generated 7.8 score with AC:L/C:H); no public exploit code or CISA KEV entry was identified at time of analysis.
A missing read memory barrier in the Linux kernel's in-kernel NFS server (nfsd) netlink status-dump path lets a low-privileged local user (CVSS PR:L) on weakly-ordered CPUs such as ARM or POWER read RPC status fields that have been concurrently modified by nfsd_dispatch, because the hand-rolled seqcount retry in nfsd_nl_rpc_status_get_dumpit() never orders its field reads ahead of the second counter check. The result is information disclosure of inconsistent or torn status data (CVSS C:H/I:N/A:H per the supplied vector), not memory corruption or code execution, and the reordering does not manifest on strongly-ordered x86/x86_64 hardware, which eliminates practical exploitability there. No public exploit code has been identified at time of analysis and there is no KEV confirmation of active exploitation, though our independent assessment scores the issue lower (AV:L/AC:H/PR:L/UI:N/S:U/C:L/I:N/A:L) than the supplied 7.1 base score because of the architecture restriction, the local-only vector, and the narrow timing race required against live NFS traffic.
Linux kernel NFS/RDMA servers using svcrdma accept RPC-over-RDMA Write and Reply chunks whose segcount field is zero, because xdr_check_write_chunk() only rejects segcount values above rc_maxpages and xdr_inline_decode(stream, 0) succeeds without advancing the cursor, causing pcl_alloc_write() to link a struct svc_rdma_chunk with ch_segcount == 0 onto the rc_write_pcl or rc_reply_pcl parsed-chunk lists even though RFC 8166 assigns no meaning to a chunk describing no remote buffer. The vulnerable path is reachable only on NFS servers configured with the RDMA transport over InfiniBand, RoCE, or iWARP, so an attacker must already be an RDMA peer able to send Write or Reply chunks to the server, and the independent assessment rates the realistic impact as availability-only (CVSS 3.1 AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:N/A:H) rather than the 9.1 input score, because an earlier patch in the same series already made pcl_for_each_segment() safe for ch_segcount == 0. No public exploit code and no confirmed active exploitation are identified at time of analysis, and vendor-released patches are available in 6.12.111, 6.18.50, 7.2.4, and mainline 7.3-rc1.
The Linux kernel's fsl-imx8mq-usb PHY driver leaks the Type-C (TCA) switch reference when probe fails after imx95_usb_phy_get_tca() has already succeeded, because the only release path lived in the driver's .remove() callback, which does not execute on probe failure. Only systems running NXP i.MX8MQ/i.MX95 hardware with the Type-C switch path active are affected, and triggering the bug requires high local privileges to force repeated probe failures; a single probe failure produces only a one-time leak of limited availability impact, and no public exploit has been identified at time of analysis (no KEV confirmation). The headline CVSS 7.4 with PR:N and full C/I/A:H is inflated relative to the assessed CVSS:3.1/AV:L/AC:H/PR:H/UI:N/S:U/C:N/I:N/A:L, which reflects a privileged, hardware-dependent, low-impact resource leak rather than a remotely exploitable compromise.
Failed connection acceptance in the Linux kernel's svcrdma transport (server-side NFS-over-RDMA) can corrupt reference-count and XArray registration state, leaking transport memory and mismatched device references. The bug is reachable only on hosts running the kernel NFS server over an RDMA fabric (InfiniBand or RoCE) where svc_rdma_accept() takes its error path before rpcrdma_rn_register() succeeds, and no public exploit has been identified at time of analysis. The assessed impact is primarily availability (repeated failed accepts leak svcxprt_rdma and its net/ns_tracker, leading to memory exhaustion) with a limited integrity effect from an erroneous erase of another caller's slot-0 entry; the CVSS vector carries PR:N, so no authentication is required, but attacks must originate from the adjacent RDMA fabric.
Use-after-free in the Linux kernel's server-side SUNRPC RPCSEC_GSS credential decoder (svcauth_gss_decode_credbody) leaves the pooled per-request credential (svcdata->clcred in gss_svc_data) holding a dangling gc_ctx.data inline pointer into a released request XDR page while gc_ctx.len still carries a stale length from an earlier request, so a downstream length-driven consumer such as gss_svc_searchbyctx() walks freed memory. Exposure is limited to hosts running an NFS server with RPCSEC_GSS/Kerberos security enabled (exports mounted with sec=krb5, krb5i, or krb5p); systems using only AUTH_SYS or running no NFS server are not affected. A remote, unauthenticated attacker who can reach the GSS-enabled service and craft a credential body that fails XDR decode can trigger the fault, but reliability is limited because the dangling state depends on residue from a prior request and on the request pages being released, so denial of service with limited information disclosure is the realistic outcome rather than assured code execution; no public exploit identified at time of analysis, and vendor patches are available for the affected stable trees.
A stack-based buffer overflow in Altera's Trusted Firmware port for SoC FPGA devices (the SoCFPGA branch through socfpga_v2.14.0) lets a local attacker who already controls the non-secure rich OS - root or kernel on the HPS ARM cores - corrupt the secure monitor's stack via the SMC interface and potentially escalate across the trust boundary into the secure world. The assessed vector (CVSS:3.1/AV:L/AC:L/PR:H/UI:N/S:C/C:L/I:H/A:H) confirms this is an access-gated privilege/trust-boundary escalation requiring an existing high-privileged foothold, not a remotely reachable default-configuration flaw; impact to integrity and availability is high and scope is changed, but confidentiality impact is limited. No public exploit code has been identified at time of analysis and the issue is not confirmed as actively exploited (not in CISA KEV).
Heap over-read in DCMTK through 3.7.0 allows unauthenticated remote attackers to crash applications or leak adjacent heap memory by supplying a crafted DICOM instance whose declared NumberOfFrames exceeds the actual PixelData buffer length. The flaw resides in the ConcatenationLoader module (dcmfg) and is triggered when loading multi-frame DICOM concatenations; exploitation requires an ingestion path that processes untrusted instances (e.g., a DICOM network storage service), as offline viewers typically require a user to open the file. No public exploit code is identified at time of analysis, and a vendor patch is available via upstream commits.
Denial of service in NodeRedis redis-parser through 3.0.0 lets a hostile or hijacked Redis endpoint crash the Node.js client process by returning a crafted RESP multi-bulk header whose declared length exceeds 2^32-1, triggering an uncaught RangeError during parsing. The flaw is reachable by unauthenticated remote attackers (CVSS 4.0 AV:N/AC:L/AT:N/PR:N/UI:N) but its impact is strictly availability (VA:H) - no confidentiality, integrity, or scope impact - so a malicious server, a compromised legitimate Redis instance, or an on-path injector in the RESP response stream can only take down the parsing process. Rated 8.7 (HIGH) by CVSS 4.0, with no public exploit identified at time of analysis and no confirmed vendor-released patch, though the issue is publicly tracked upstream.
Command injection in rpm enables an attacker to run arbitrary shell commands as the invoking user when a victim installs, rebuilds, or otherwise processes a malicious source RPM (.src.rpm) that contains a %() macro construct in a source or spec file basename. The flaw is local and requires user interaction, with impact confined to the privileges of the user running rpm or rpmbuild, typically non-root; build farms and CI/CD pipelines that automatically handle untrusted SRPMs are the highest-risk targets. Red Hat CPEs list RHEL 6 through 10 and Red Hat Hardened Images, and no public exploit was identified in the provided intelligence; monitor the vendor advisory for exploitation and patch updates.
Local attackers can achieve arbitrary code execution by tricking a user or automation into processing a crafted manifest file with 'rpm -q -p', due to unexpected macro expansion of manifest entries in RPM before they are opened. This command injection flaw (CWE-78) affects Red Hat Enterprise Linux 6 through 10 and Red Hat Hardened Images, with a CVSS 3.1 base score of 7.8 (HIGH, AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H); exploitation requires user interaction but no privileges, and impact is limited to the privileges of the rpm process. No public exploit code or active exploitation has been identified at time of analysis.
Privilege escalation in Rapid7 Velociraptor allows an authenticated user holding only the 'investigator' role to execute arbitrary VQL statements with administrator privileges on the Velociraptor server by populating the internal 'compiled_collector_args' field when submitting a hunt-creation/scheduling API request. Because the server accepts this internal field from the caller, the attacker-supplied pre-compiled VQL is trusted and the ACL checks that normally gate artifact compilation are skipped (CVSS 9.9, AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H). Exploitation requires a valid investigator-or-higher account that can schedule hunts, making this an insider/privilege-escalation path rather than an unauthenticated remote compromise; no public exploit identified at time of analysis, and no confirmed actively exploited (CISA KEV) status was provided. The vendor advisory is published at https://docs.velociraptor.app/announcements/advisories/cve-2026-19072/.
An out-of-bounds write in GIMP's GIMPressionist plug-in allows an attacker to corrupt memory by convincing a user to load a malicious preset file, potentially leading to a crash or arbitrary code execution. The flaw is an unauthenticated local attack (CVSS PR:N) that requires user interaction (UI:R) to open a crafted preset file, and it affects GIMP as packaged in Red Hat Enterprise Linux 6 through 10 plus likely upstream GIMP versions. No public exploit has been identified at time of analysis, and the overall risk is assessed as genuine but not urgent due to the requirement for user interaction with a crafted file.
Heap-based buffer overflow in the Nanomsg WebSocket transport (versions 0.5-beta through 1.x before 1.2.3) allows remote attackers to corrupt heap memory by sending an oversized Sec-WebSocket-Version header during the WebSocket handshake, which is processed before any authentication or user interaction takes place. Exploitation requires that the target application actually enable and expose nanomsg's ws transport; TCP- or IPC-only deployments are unaffected. The flaw yields high integrity and availability impact, and while reliable code execution depends on heap layout and platform, no public exploit code has been identified at time of analysis and the vulnerability is not confirmed as actively exploited (not in CISA KEV).
Denial of service in the MuJS embeddable JavaScript interpreter (git revision e892c9fdb) allows remote attackers to crash the hosting process by supplying JavaScript that builds an excessively large numeric array index, which jsR_isindex() in jsrun.c converts from floating point to integer without validating the value fits in the target integer range. Per the assessed CVSS vector (AV:N/AC:L/PR:N/UI:N/S:U/C:N/I/A:H), the attack is network-reachable and unauthenticated, but the impact is availability-only: no confidentiality or integrity loss, with exploitability gated by whether the embedding application evaluates untrusted JavaScript. Exploitation requires the host to feed attacker-controlled script to the embedded interpreter (for example, opening a document or file that MuJS parses), and no public exploit code or CISA KEV entry was identified at time of analysis; EPSS is 0.15% (4th percentile), indicating low near-term exploitation likelihood. The upstream project has published a fix commit, but no tagged patched release has been independently confirmed.
Crafted DNS traffic parsed by the nDPI 5.1.0 deep-packet-inspection library can terminate the hosting process wherever that code runs on a strict-alignment CPU (ARM, MIPS, SPARC and similar) or in an UndefinedBehaviorSanitizer/fuzzing build, because the DNS dissector and the serializer deserialization path dereference byte buffers at arbitrary packet offsets through uint16_t/uint32_t/uint64_t pointers without any alignment check. The flaw is remotely reachable and unauthenticated (CVSS 3.1 AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H, 7.5), but impact is limited to availability - process termination, no confidentiality or integrity loss - and the decisive precondition is the build/CPU target: mainstream x86-64 release builds silently tolerate the misaligned load and generally will not crash. It is not listed in CISA KEV and no public exploit identified at time of analysis, although the upstream fix PR does ship fuzz corpus inputs (issue3213_trigger1-3) that reproduce the fault, and EPSS data was not available for this record.
Denial of service in hiredis post-v1.5.0 (commit 29ea279) occurs when the RESP aggregate parser performs an uncontrolled memory allocation based on an oversized element or length count supplied in a Redis server reply. A hiredis-based client that connects to a malicious or compromised Redis server, or whose unencrypted connection is subject to man-in-the-middle redirection, can be driven into memory exhaustion and crash or unresponsiveness; no authentication on the hiredis side and no user interaction are required. The provided data does not include a CISA KEV entry, and no public exploit identified at time of analysis, so exploitation should be assessed as requiring an attacker-controlled or untrusted Redis endpoint rather than as confirmed active exploitation.
Type confusion in the Perl DBI module's fetchrow_hashref implementation allows an application that assigns a non-string scalar (integer/IV or float/NV) to the FetchHashKeyName attribute and then calls fetchrow_hashref to trigger an out-of-bounds memory read, typically crashing the Perl process (denial of service) and, in principle, exposing adjacent heap memory in error output. Affected are DBI releases before 1.654 on any platform where the module is installed; the CVSS 3.1 vector (AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H, 7.5) rates this unauthenticated and remotely triggerable, though independent assessment raises attack complexity to AC:H because exploitation requires the hosting application to propagate attacker-controlled data into a statement-handle attribute that is normally set by the application itself. No public exploit code has been identified at time of analysis, no confirmed active exploitation (not in CISA KEV), and EPSS is low at 0.21% (10th percentile), reflecting the narrow precondition of application-specific attribute misuse.
Denial of service in NanoSVG at commit 239e102ec affects applications that parse attacker-supplied SVG documents, where a crafted arc command with extreme radius values drives the arc delta-angle calculation to NaN and the subsequent unchecked NaN-to-int conversion causes undefined behavior and process termination. The published CVSS vector is 7.5 (AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H), so the unauthenticated network attack path is constrained by the requirement that the target actually process a malicious SVG, and manual SVG opening can effectively require user interaction; EPSS is 0.16% and no public exploit identified at time of analysis.
Heap out-of-bounds write in QuickJS's bytecode deserializer (JS_ReadFunctionTag) lets a crafted serialized blob corrupt heap memory, giving an attacker the potential for arbitrary code execution in the context of the host process. The flaw is only reachable when the embedding application calls JS_ReadObject with the JS_READ_OBJ_BYTECODE flag on attacker-controlled input - not the default pattern of executing JavaScript source - so exposure is confined to plugin, cache, or IPC loaders that treat serialized bytecode as untrusted data; where that path is exposed, the flaw is remotely triggerable by an unauthenticated attacker (CVSS vector PR:N) with a single malformed blob and no user interaction. Public exploit code has not been identified at time of analysis (EPSS 0.17%, 5th percentile), and while the vendor score is 9.8 with AC:L, an independent assessment rates attack complexity as high (CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:H) because exploitation depends on a non-default deserialization path in the hosting application.
Denial of service in PHP: attacker-influenced input passed into a recursive routine can exhaust process resources and terminate the PHP worker or interpreter (CVSS 3.1 base 7.5, AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H), with impact confined to availability and no data exposure or code execution. The flaw is network-reachable, unauthenticated, and low-complexity per the vector, so any deployment that feeds untrusted remote input into the affected recursive code path is exposed, while deployments that only pass trusted input are not. No public exploit code or confirmed active exploitation was identified at time of analysis, and the upstream PHP security advisory GHSA-rgrp-mwpx-f6rm is the authoritative source for the exact affected and fixed versions.
Crafted MAT4 (MATLAB v4) files cause an integer overflow in libsndfile 1.2.2's mat4_read_header() parser, crashing the process and producing a denial of service. The flaw is network-reachable without authentication (CVSS 3.1 PR:N), but our assessment requires a victim to open the file or a backend service to ingest untrusted audio (UI:R); impact is limited to availability, with no code execution or data disclosure. No public exploit code and no vendor-released patch are confirmed at time of analysis.
Denial of service in NanoSVG (commit 239e102ec) where a crafted SVG document carrying oversized geometry coordinates drives the rasterizer's fixed-point-scaled edge values past the range of a 32-bit int during nsvg__addActive(), producing undefined behavior that can terminate the process. The issue is rated CVSS 3.1 7.5 (AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H), i.e. unauthenticated and low-complexity, but availability-only: there is no code execution, data disclosure, or integrity impact per the assessed vector. Exploitation requires the embedding application to actually rasterize attacker-supplied SVG content rather than merely parse it, so tools that only handle locally trusted assets face a local-input scenario; no public exploit code was identified at time of analysis and the flaw is not confirmed actively exploited (not in CISA KEV), though the vendor issue tracker reference (GitHub issue 292) makes root-cause details publicly discoverable.
TLS client-authentication bypass in the RabbitMQ trust-store plugin lets a remote unauthenticated attacker whose forged self-signed certificate carries the issuer DN and serial number of any whitelisted certificate pass mutual-TLS authentication. Only deployments that have the rabbitmq_trust_store plugin enabled and wired in as the TLS verify_fun for a client-certificate listener are affected, and an attacker must first learn or guess at least one whitelisted certificate's {Issuer, Serial} pair, which is non-secret and is exposed by the plugin's list command, logs, or any copy of the certificate. All RabbitMQ server releases prior to 3.13.15, 4.0.20, 4.1.11, 4.2.6, and 4.3.0 are vulnerable; vendor-released patches exist in those versions, and no public exploit identified at time of analysis.
Memory-exhaustion denial of service in RabbitMQ's Web-MQTT plugin (rabbitmq_web_mqtt) lets an unauthenticated remote attacker crash a broker node by sending a single WebSocket frame that is highly compressed and inflates to gigabytes in memory. Affected are RabbitMQ server releases prior to 3.13.15, 4.0.20, 4.1.11, 4.2.6 and 4.3.0 whenever the Web-MQTT plugin is enabled and its listener (TCP 15675/15676) is network-reachable; the attacker only needs to negotiate RFC 7692 permessage-deflate during the upgrade and then send a zlib bomb, since decompression happens in the connection process before any MQTT CONNECT credentials are evaluated. No public exploit code was identified at time of analysis, exploitation requires the non-default plugin to be enabled, and the impact is availability-only (C:N/I:N/A:H) against the affected node. Vendor-released fixes are available in 3.13.15, 4.0.20, 4.1.11, 4.2.6 and 4.3.0.
Denial of service in RabbitMQ Server (versions prior to 4.3.0, 4.2.6, 4.1.11, 4.0.20, and 3.13.15) allows an authenticated AMQP 0-9-1 client with publish permission to exhaust broker memory by declaring an excessive content-header BodySize and streaming message fragments without completing assembly. The unvalidated uint64 BodySize bypasses the max_message_size check, which only runs at assembly completion, causing unbounded memory accumulation in the reader process until the memory alarm triggers or the node runs out of memory, degrading all publishers cluster-wide. No public exploit code has been identified at time of analysis, and vendor patches are available.
Denial of service in RabbitMQ versions prior to 4.2.7 and 4.3.1 allows any authenticated AMQP 0-9-1 client to crash the entire Erlang VM, taking down all vhosts and connections, by exhausting the atom table with roughly one million requests. The attack exploits the direct reply-to feature: queue names or routing keys beginning with 'amq.rabbitmq.reply-to.' cause an ETF-encoded pid node name to be passed to binary_to_atom without using binary_to_existing_atom, and because atoms are never garbage collected, the try/catch cannot reclaim them. Exploitation requires a valid AMQP connection (any credential, any permission level) and the ability to send the large number of operations; no public exploit identified at time of analysis.
Unauthenticated remote attackers can crash any RabbitMQ node with the AMQP 1.0 listener enabled by sending a single ~19-byte frame to TCP port 5672, causing the Erlang VM to exhaust heap memory and terminate, which takes down all tenants and protocols on that node. The flaw is an unbounded-allocation bug (CWE-770) in the AMQP 1.0 framing code that is reached before authentication completes, so no credentials or user interaction are required; only network reachability to the broker's AMQP 1.0 port is needed. Affected releases are RabbitMQ prior to 3.13.15, 4.0.20, 4.1.11, and 4.2.6, which contain the vendor-released fix; no public exploit code has been identified at time of analysis, and the issue is scored high severity (CVSS 8.2 v4.0 / assessed CVSS 3.1 7.5) with availability-only impact.
The RabbitMQ management plugin (rabbit_mgmt_cors.erl) reflects an attacker-controlled Origin header straight back in Access-Control-Allow-Origin while also returning Access-Control-Allow-Credentials: true, so when an operator has configured cors_allow_origins = "*" a hostile web page can make a signed-in administrator's browser issue authenticated, state-changing requests to the management HTTP API. Exploitation is conditional rather than automatic: it requires the non-default wildcard CORS setting, a live cached HTTP Basic-auth administrator session, and the administrator being lured to the attacker's page (user interaction), and no public exploit code has been identified at time of analysis; assessed severity is CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:N. The issue affects the management plugin in RabbitMQ releases prior to 3.13.17, 4.0.22, 4.1.13 and 4.2.6.
RabbitMQ brokers that use the OAuth2 plugin can silently fall back to disabled TLS certificate verification when no CA bundle can be located, allowing an attacker already positioned as a network man-in-the-middle to forge the JWKS response and inject arbitrary JWTs that the broker will accept as authentic. The flaw affects RabbitMQ server versions prior to 3.13.15, 4.0.20, 4.1.11, 4.2.6, and 4.3.0, and exploitation requires four conditions to align: the OAuth2 plugin must be in use, no cacertfile may be configured, the host or container's OS CA bundle must be empty or unreadable, and the attacker must hold a MitM position on the path to the identity provider's JWKS endpoint. Unauthenticated remote attackers are the relevant threat actor per the CVSS vector (PR:N), but the precondition stack and the required network position make this a heavily gated authentication-bypass rather than a blanket internet-exploitable critical; no public exploit code and no confirmed active exploitation were identified at time of analysis.
Stored cross-site scripting in the RabbitMQ management UI allows an attacker who can obtain a TLS client certificate signed by a CA the broker trusts to embed JavaScript in the certificate's Subject DN, which is rendered unescaped in the connection detail page and executes in an administrator's browser session when that connection is viewed. Success enables full administrative account takeover within the management UI - creating users, exporting definitions, and performing any action the viewing admin can perform. Exploitation is heavily gated: the TLS listener must be configured with ssl_options.verify = verify_peer (not the default), the certificate must be signed by a CA in the broker's trust store (self-signed certificates do not work), and an administrator must open the affected connection page; the management UI's CSP includes 'unsafe-inline', so inline script is not blocked. Patched versions are 3.13.15, 4.0.20, 4.1.11, and 4.2.6; no public exploit code and no confirmed active exploitation were identified at time of analysis.
Use-after-free in librsvg during rendering of crafted SVG documents with nested XInclude directives and duplicate entity declarations allows unauthenticated local attackers to crash or potentially execute code in applications that use the library. Tracked as CVE-2026-96889 and reported by Red Hat, the flaw affects librsvg as shipped in Red Hat Enterprise Linux 6, 7, 8, 9, and 10. Exploitation requires user interaction (UI:R) because a victim must open, preview, or thumbnail a malicious SVG; the CVSS 3.1 score is 7.8. No public exploit code was identified at time of analysis, and independent assessment indicates that arbitrary code execution is unproven-only denial of service or crash is reliably reachable.
Deeply nested JSON arrays and objects drive JVM heap exhaustion in Typelevel Jawn versions before 1.7.0, where the thrown java.lang.OutOfMemoryError is a fatal Scala/JVM error that neither scala.util.Try nor cats.effect.IO routinely intercepts, turning untrusted JSON into a denial-of-service condition. The assessed impact is availability-only (CVSS:3.1 7.5, AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H), so remote unauthenticated attackers need only reach an endpoint that feeds attacker-controlled input into a pre-1.7.0 Jawn parse call, directly or transitively through circe; services that only parse trusted or internal JSON are not exposed. No public exploit code has been identified at time of analysis and the issue is not listed in CISA KEV, but this is a genuine, middle-tier priority for any internet-facing JSON API on an affected version; fixed in 1.7.0, which introduces a configurable depth cap.
Denial of service in the Jawn JSON parser for Scala (all artifacts prior to 1.7.0) lets a remote, unauthenticated attacker exhaust CPU by feeding JSON through the streaming AsyncParser in many tiny chunks, forcing repeated rescans of a single incomplete token. Exploitation is bounded by two preconditions: the target must parse attacker-influenced input incrementally via AsyncParser rather than parsing a fully buffered document, and the attacker must be able to influence chunk boundaries so that one token is split across many small reads - hence the assessed CVSS 3.1 vector AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:N/A:H, an availability-only (HIGH) impact with no confidentiality or integrity effect. No public exploit code and no confirmed active exploitation were identified at time of analysis; the vendor fixed the issue in version 1.7.0.
Denial of service in JLine's built-in less viewer affects all 3.x releases from 3.0.0 before 3.30.15 and 4.x releases before 4.3.1: a user-supplied search or display-filter pattern containing nested quantifiers is handed straight to Java's backtracking regex engine and applied repeatedly across file content, so a pattern run against non-matching lines can consume excessive CPU and block the session thread indefinitely. The practical risk profile is multi-user Telnet or SSH deployments that embed the less viewer, where repeated sessions can exhaust a bounded worker pool and deny service to other users; the independent assessment scores this as availability-only (CVSS 3.1 A:H, no confidentiality or integrity impact) and notes that an interactive session foothold and viewable file content are prerequisites, which the published vector's PR:N rating does not reflect. No public exploit code and no confirmed active exploitation were identified at time of analysis, so exploitation is opportunistic rather than observed. The fix is the SafeRegex wrapper plus regex timeouts shipped in 3.30.15 and 4.3.1.
Denial of service in JLine3's built-in grep command (org.jline:jline-builtins) lets anyone who can supply a search pattern hang a worker thread indefinitely via catastrophic regular-expression backtracking. The builtin - reachable in applications that embed JLine3 and expose the shell to remote or interactive users - rewrites non-line-regexp patterns as .*pattern.* before compiling them with Java's backtracking regex engine, so a short input such as (a+)+b tested against non-matching text causes exponential CPU consumption (CWE-1333; CVSS 3.1 base 7.5 upstream, while the assessed vector requires low privileges, PR:L). Impact is availability-only (C:N/I:N/A:H) with no confidentiality or integrity loss; publicly available exploit code exists in the form of the advisory's two-line PoC, but the issue is not listed in CISA KEV and no confirmed active exploitation has been reported. Vendor-fixed releases are available: 3.30.15 on the 3.x line and 4.3.1 on the 4.x line.
Local privilege escalation to root in Flatpak before 1.18.1 lets an unprivileged user holding an active local session abuse the revokefs writer used by flatpak-system-helper: by opening two concurrent revokefs sessions and planting a symlink from one session directory into the other, the attacker retains a file descriptor into a peer session and keeps modifying its files after the system helper has already validated and imported them. The practical payoff is tampering with ostree commit objects in the system repository after signature verification, which translates into root-controlled writes to attacker-chosen paths. Exploitation requires local access and winning a TOCTOU race against the dual-session setup (CVSS AV:L/AC:H/PR:N), and no public exploit code has been identified at time of analysis; the issue is not confirmed as actively exploited (not in CISA KEV).
A heap buffer overflow in BusyBox's romfs filesystem image parsing path can be triggered when the utility processes an attacker-crafted romfs image whose volume ID metadata is read with an unbounded strlen, producing an out-of-bounds heap write (CWE-787) rated CVSS 3.1 base 7.3. Because the attack vector is local (AV:L), the malicious image must be delivered to the target system - for example via removable media, a downloaded archive, a firmware/container image, or a package - and some process or user must then parse or mount it (UI:R); it is not remotely triggerable over a network. Under the assessed vector (AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H) no privileges are required, and successful corruption could yield code execution in the context of the parsing process with high impact to confidentiality, integrity, and availability; no public exploit or POC was identified in the supplied intelligence and there is no confirmed active exploitation.
Pre-authentication heap memory corruption in BusyBox's built-in TLS implementation can be triggered by a remote, unauthenticated attacker who sends a single crafted ClientKeyExchange handshake message to a BusyBox process acting as a TLS server, corrupting the heap and crashing the process (CVSS 3.1 base 7.5, availability-only with no confidentiality or integrity impact rated). Exploitation requires a BusyBox binary compiled with TLS support and reachable in a server role, a deployment pattern that is the exception rather than the norm since BusyBox TLS is most often used outbound (for example by wget https), which sharply narrows the realistically exposed population. No public exploit code or confirmed active exploitation (CISA KEV) was identified in the available intelligence; the issue is tracked by Red Hat via its CVE page and Bugzilla, with no public exploit identified at time of analysis.
Path traversal in Flatpak's `flatpak build-init` command allows an attacker-controlled SDK container to write arbitrary files outside the developer's working directory when that developer builds against the SDK using both the `--writable-sdk` and `--sdk-extension` flags. The root cause (CWE-22) is that the SDK's declared extension-point `directory` value is resolved through a path helper that does not reject `..` components, so a malicious SDK can steer writes outside the intended build root. The independently assessed vector (CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H) is a local, user-interaction-dependent action rather than the remote 9.8 originally submitted; no public exploit code or confirmed active exploitation has been identified at time of analysis, and the attacker must convince the developer to consume a hostile SDK, which severely limits real-world exposure.
Message-filtering bypass in xdg-dbus-proxy before 0.1.9 lets a malicious or already-compromised sandboxed application - most notably a Flatpak app, but also deployments under frameworks such as Firejail - escape its confinement and execute arbitrary code on the host by attaching a forged reply serial number to non-reply D-Bus messages. The flaw carries a CVSS 4.0 base score of 8.7 (AV:N/AC:L/PR:N/UI:P, with high confidentiality, integrity and availability impact scoped to the sandbox-escape boundary) and is rated a genuine high-priority sandbox-escape issue rather than a CVSS-inflated paper risk; the practical precondition is that an attacker must first get code running inside a sandbox that routes its session bus through the vulnerable proxy, typically by tricking a user into installing or launching a hostile or trojanized app. No confirmed active exploitation (CISA KEV) and no public exploit code were identified at time of analysis; vendor-released patch: 0.1.9.
Apache Tomcat Native 1.3.0 through 1.3.8 and 2.0.0 through 2.0.15 can, under a narrow timing window, down-grade the client-certificate verification requirement for connections, weakening mutual-TLS authentication for deployments that enforce client-cert auth through the APR/OpenSSL-based connector. The flaw is a CWE-366 race condition within a thread (CVSS 3.1 base 7.4, AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:N), so an unauthenticated network attacker must win a non-deterministic race to succeed, making exploitation probabilistic rather than guaranteed. Only Tomcat instances that actually use Tomcat Native and are configured for client-certificate verification are exposed; no public exploit identified at time of analysis, and no CISA KEV confirmation of active exploitation. Vendor-released fixes are available in 2.0.16 and 1.3.9.
Insecure-by-default OpenSSL options in Apache Tomcat Native (versions 2.0.0 through 2.0.15 and 1.3.0 through 1.3.8) silently weaken the TLS layer used by the APR/OpenSSL connector, enabling insecure client-initiated renegotiation, disabling the Extended Master Secret extension, ignoring unexpected EOF (truncation), and permitting key exchanges without (EC)DHE forward secrecy. The vendor scored this 9.1 (CVSS:3.1/AV:N/AC:L/PR:N/UI:N/C:H/I:H/A:N), but our independent assessment rates it substantially lower - CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:L/A:N - because meaningful abuse requires an on-path/man-in-the-middle attacker who can actively position themselves between client and server, and because the four options are only applied when the deployment has not explicitly overridden them. Only Tomcat instances configured with the Apache Tomcat Native (APR) TLS connector at an affected version are exposed; JSSE/NIO-based connectors are unaffected, and no public exploit code or CISA KEV entry was identified at time of analysis. Vendors/users should upgrade to Tomcat Native 2.0.16 or 1.3.9.
Denial of service via JVM crash affects Apache Tomcat Native versions 2.0.0 through 2.0.15 and 1.3.0 through 1.3.8, where a buffer over-read (CWE-126) during the TLS handshake lets a malicious client terminate the server process, rated CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H. The flaw is reachable remotely and unauthenticated with no user interaction, but only where the APR/OpenSSL-based native TLS connector is actually deployed and handling TLS connections; Tomcat instances using the pure-Java JSSE/NIO connector are not affected. Impact is confined to availability - no confidentiality or integrity loss, no data exposure and no code execution - and no public exploit was identified at time of analysis, with no CISA KEV entry.
WebSocket message smuggling in Apache Tomcat allows a remote, unauthenticated attacker to inject or reorder messages within a compressed WebSocket stream, corrupting application-level message boundaries and undermining integrity of data exchanged over the connection. The flaw affects the 11.0.0-M1 through 11.0.25, 10.1.0-M1 through 10.1.59, and 9.0.0.M1 through 9.0.121 release lines, plus the end-of-support 8.5.0-8.5.100 and 7.0.56-7.0.109 branches, but it only triggers when the per-message-deflate compression extension has been negotiated on the WebSocket connection; deployments that serve only standard HTTP or that run WebSocket endpoints without compression are not exposed. There is no confirmed active exploitation (CISA KEV does not list this issue), no public exploit code identified at time of analysis, and CISA's SSVC assessment rates exploitation as 'none', making this a genuine but conditionally-gated integrity flaw rather than a drop-everything emergency.
HTTP/2 request smuggling in Apache Tomcat can mix up request headers between concurrent streams, allowing unauthenticated remote attackers to potentially read or alter other users' request data when the connector has HTTP/2 explicitly enabled. The flaw is a regression from the fix for CVE-2026-41293 and affects only Tomcat 11.0.22 through 11.0.25, 10.1.55 through 10.1.59, and 9.0.118 through 9.0.121; exploitation is rated high complexity (AC:H) with limited confidentiality and integrity impact (C:L/I:L). No public exploit code has been identified at time of analysis, and the independent assessment rates this as a moderate-priority, genuine protocol-layer flaw rather than a critical emergency.
Authentication can succeed when it should have been rejected in Apache Tomcat deployments that use CLIENT_CERT (mutual-TLS) authentication with soft fail disabled, allowing network attackers to reach resources that the mutual-TLS configuration was meant to protect. The affected range is broad - 9.0.92 through 9.0.121, 10.1.22 through 10.1.59, and 11.0.0-M14 through 11.0.25 - but the flaw is gated on two non-default conditions (the CLIENT_CERT authenticator in use and soft fail turned off) and, per the vendor, only manifests in some scenarios, so it is not universally reproducible even on correctly configured vulnerable instances. No public exploit code has been identified at time of analysis, and no evidence of active exploitation is present; the attacker needs network reachability to the mTLS-protected listener.
Denial of service in Apache Tomcat's asynchronous WebSocket write path allows remote, unauthenticated attackers to exhaust server resources by triggering lost write timeouts, so non-blocking WebSocket sends are never released after their effective lifetime. All maintained branches are affected - 11.0.0-M1 through 11.0.25, 10.1.0-M1 through 10.1.59, and 9.0.0.M1 through 9.0.121, plus the end-of-life 8.5.0-8.5.100 and 7.0.43-7.0.109 lines - although exposure is bounded to instances whose deployed applications actually perform asynchronous (non-blocking) WebSocket writes; servers with no WebSocket endpoints or only synchronous writes are not exposed. The flaw is rated CVSS 7.5 (AV:N/AC:L/PR:N/UI:N, availability-only) and CISA's SSVC assessment records no observed exploitation, with no public exploit code identified at time of analysis; the primary risk is sustained resource pressure that degrades or halts the affected connector rather than data compromise.
Apache Tomcat 11.0.19 through 11.0.25, 10.1.53 through 10.1.59, and 9.0.116 through 9.0.121 can cause an unrelated user's request to fail when a malformed request triggers incomplete cleanup of request-processing state on a reachable HTTP connector. No authentication or user interaction is required, but the effect is fundamentally gated by a race window - the advisory itself states the failure occurs only 'depending on timing' - so it is non-deterministic and cannot be reliably triggered on demand, and the realistic worst case is intermittent failure of another user's individual request rather than sustained denial of service, data disclosure, or code execution. No public exploit code and no confirmed active exploitation were identified at time of analysis; the 7.3 base score reflects low impact across confidentiality, integrity, and availability, though our assessment rates attack complexity as High rather than Low.
Denial of service in Apache Tomcat allows an attacker to pin an AJP processing thread through a single AJP request, exhausting the connector's limited thread pool (CWE-770, allocation of resources without limits or throttling). The affected range is broad: 11.0.0-M1 through 11.0.25, 10.1.0-M1 through 10.1.59, 9.0.0.M1 through 9.0.121, plus the end-of-life branches 8.5.0-8.5.100 and 7.0.0-7.0.109, which remain affected even though they are no longer supported. There is no public exploit identified at time of analysis, no KEV listing, and no EPSS score in the input; the independent assessment rates the issue CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H (availability-only impact), but exploitation requires the AJP connector to be enabled and network-reachable, which has not been the default configuration since Tomcat 9.0.31.
Uncontrolled resource consumption in Apache Tomcat's WebSocket close-message handling allows unauthenticated remote attackers to exhaust server resources and degrade availability on instances that expose WebSocket endpoints. Per the assessed vector (CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L), the impact is limited to availability (Low) with no confidentiality or integrity effect, and no user interaction is required. Affected releases are 11.0.0-M5 through 11.0.25, 10.1.8 through 10.1.59, 9.0.74 through 9.0.121, plus the end-of-life 8.5.88 through 8.5.100 line; fixes are available in 11.0.26, 10.1.60, and 9.0.122. There is no confirmed active exploitation (not in CISA KEV) and no public exploit identified at time of analysis, but the low attack complexity and unauthenticated network reachability make prompt patching advisable for any deployment using WebSockets.
Apache Tomcat's HTTP/2 connector harbors a race condition (CWE-362) that allows unauthenticated remote attackers to inject trailer fields into another user's concurrent HTTP/2 request, resulting in high confidentiality and integrity impact without affecting availability. Affected versions are 11.0.0-M1 through 11.0.25, 10.1.0-M1 through 10.1.59, 9.0.39 through 9.0.121, and the end-of-life 8.5.59 through 8.5.100; exploitation is conditional on winning a concurrency race and only applies when the HTTP/2 upgrade protocol is enabled and actively serving traffic. No public exploit code or active exploitation has been identified at time of analysis, and the overall threat is considered credible but conditional.
Authentication bypass in Apache Tomcat allows unauthenticated remote attackers to reach WebSocket endpoints that are supposed to be protected by container-managed security constraints, simply by referencing the endpoint under an alternate name so the constraint is never matched. Exploitation requires only that the application expose one or more WebSocket endpoints and rely on container-level (web.xml security-constraint) authentication or authorization rules to guard them; no credentials, no user interaction, and no non-default configuration beyond that are needed, and per the CVSS vector (PR:N) the attack is unauthenticated and network-reachable. Affected lines are Tomcat 11.0.0-M1 through 11.0.25, 10.1.0-M1 through 10.1.59, 9.0.0.M1 through 9.0.121, plus end-of-support 8.5.0-8.5.100 and 7.0.43-7.0.109; no public exploit code has been identified at time of analysis and Apache has released fixed versions 11.0.26, 10.1.60 and 9.0.122.
Cross-application authentication realm leakage in Apache Tomcat's Jakarta Authentication integration causes requests to be validated against the realm established by whichever co-hosted web application authenticated first, breaking per-application authentication boundaries. The flaw (CWE-287, improper authentication) affects Tomcat 11.0.0-M1 through 11.0.25, 10.1.0-M1 through 10.1.59, and 9.0.0.M4 through 9.0.121, plus end-of-life 8.5.0 through 8.5.100. Exploitation is gated by a non-default deployment - SimpleAuthConfigProvider as the default Jakarta Authentication provider with two or more applications sharing it - and the independent assessment rates the vector CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:L/A:L (unauthenticated, high attack complexity, low impact); no public exploit identified at time of analysis, the issue is not in CISA KEV, and the vulnerable configuration is not Tomcat's default.
Code execution in GNU Emacs versions prior to 31.2 occurs when Flymake performs on-the-fly syntax checking with a language backend other than Emacs Lisp, allowing code embedded in an untrusted file to run with the privileges of the user who opened or edited it. The CVSS vector (AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H) reflects that the attack is local and requires the victim to open a malicious file, but needs no privileges and can fully compromise confidentiality, integrity, and availability of that user's session; the flaw is tracked by Red Hat as CVE-2026-96442 with a vendor score of 7.8. No public exploit code and no confirmed active exploitation were identified at time of analysis, and the risk is concentrated in developer and administration workstations where Emacs commonly opens untrusted source files from downloads, mail attachments, or repositories.
Symlink-based path traversal in the `tar` source plugin of Apache BuildStream (versions up to and including 2.8.0, running on Python < 3.12) lets a malicious source tarball write arbitrary files on the host with the privileges of the user running BuildStream, resulting in full confidentiality, integrity, and availability impact to that user's environment. The vendor-published CVSS 3.1 base score is 9.8 (AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H), but the assessed vector adds UI:R: exploitation requires the victim to fetch and extract an attacker-controlled or otherwise untrusted source tarball, and it is not exploitable against projects that use only trusted, tracked sources, since SHA256 hash pinning of tracked tarballs prevents MITM substitution. No public exploit code and no confirmed active exploitation were identified at time of analysis (CISA SSVC exploitation: none; EPSS 0.19%, 10th percentile); the vendor-released patch is BuildStream 2.8.1, and hosts running Python >= 3.12 with BuildStream >= 2.3.0 are already protected because the Python `tarfile` extraction filter blocks the symlink escape.
Denial of service in FasterXML jackson-databind allows unauthenticated remote attackers to exhaust CPU resources by submitting crafted JSON that triggers quadratic complexity in forward-reference resolution for @JsonIdentityInfo object IDs. The attack requires the application to deserialize attacker-influenced JSON into a collection or map configured with @JsonIdentityInfo; a shallow document with N forward references defined in reverse order causes O(N^2) identity comparisons during deserialization. This is a conditional availability risk-only applications using identity-enabled deserialization on untrusted input are exposed-and no public exploit code has been identified at time of analysis, though a vendor patch is available.
Memory exhaustion in FasterXML jackson-databind allows unauthenticated remote attackers to grow an internal deserializer cache without bound by submitting JSON payloads that carry ever-changing, unrecognized polymorphic type IDs. TypeDeserializerBase caches each fallback resolution keyed by the raw attacker-supplied ID string, the map has no configurable size limit, and it lives for the lifetime of the long-lived ObjectMapper, so distinct unknown IDs translate directly into permanently retained heap - the reporter measured 10,000 retained entries from 10,000 distinct unknown IDs versus a single entry for 10,000 repetitions of one unknown ID. Exploitation requires a non-default configuration (name-based polymorphism via @JsonTypeInfo(use = Id.NAME) with a defaultImpl/fallback deserializer, an endpoint that accepts attacker-influenced type IDs, and a shared application-scoped mapper), so the affected population is substantially narrower than the 7.5 base score suggests; no public exploit has been identified at time of analysis and a vendor patch is available.
Denial of service in FasterXML jackson-core 2.8.0 and later: an unauthenticated remote attacker who can supply a malformed JSON token to a parser created via JsonFactory.createParser(DataInput) causes UTF8DataInputJsonParser._reportInvalidToken() to accumulate the entire offending token into an exception message without ever applying ErrorReportConfiguration.getMaxErrorTokenLength(). The reporter measured a 20,000,109-character exception message from a 20-million-character malformed token (versus 367 characters for identical input on the InputStream path), and because the StringBuilder also incurs byte-to-char expansion and internal array doubling, scaling the payload can exhaust heap and trigger a JVM-wide OutOfMemoryError. Only the DataInput parser path is reachable, impact is availability-only, and no public exploit identified at time of analysis; the 7.5 base score captures the theoretical worst case rather than the narrow real-world reachability.
Use-after-free in Google Chrome desktop versions prior to the September 2026 stable channel update allows remote attackers to cause information disclosure and memory corruption when a victim loads attacker-controlled web content. Exploitation requires user interaction (UI:R) and a high-complexity timing or heap-state precondition (AC:H), but no authentication (PR:N); the assessed CVSS vector is AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:H, yielding a base score of 8.3. A vendor patch is available in the September 2026 stable channel update, and no public exploit code or active exploitation has been identified in the provided intelligence.
A use-after-free in Google Chrome desktop can be triggered when a victim renders attacker-controlled web content, allowing memory corruption that may lead to code execution or disclosure of sensitive in-process data. Affected are Chrome desktop builds on Windows, macOS, and Linux prior to the September 2026 stable channel update; the CVSS vector (AV:N/AC:L/PR:N/UI:R/S:C/C:H/I:H/A:H, 9.6) requires no authentication or privileges but does require the victim to open or load a malicious or compromised page, and the S:C metric indicates the attack can affect a component beyond the one that is corrupted. Vendor-released patch available; no public exploit identified at time of analysis, and the reporting originator is Ubuntu.
Ubuntu Linux is the only product attributable to CVE-2026-95319, and the sole intelligence signal available is a vendor:ubuntu report with no accompanying advisory text, version range, or affected component. No description, CVSS score, CVSS vector, CWE classification, CPE string, EPSS score, KEV status, or reference link was supplied, so the vulnerability class, attack vector, privilege requirements, and impact cannot be characterized. No public exploit code has been identified at time of analysis, but this reflects absent data rather than a confirmed absence of exploitation, and the entry must be treated as unassessable until the upstream NVD or Ubuntu security advisory is retrieved.
CVE-2026-95318 is currently an unenriched record that cannot be meaningfully risk-assessed: the only attributable data point is that it was reported by vendor:ubuntu, with no description, no CPE strings, no CWE, no CVSS score or vector, and no references supplied. There is no confirmation of what component, package, or Ubuntu release is affected, whether exploitation requires authentication or user interaction, or what the resulting impact would be. Neither confirmed active exploitation (CISA KEV) nor publicly available exploit code exists in the available data, so the exploitation status is best described as no public exploit identified at time of analysis. This analysis should be treated as provisional and re-run once Ubuntu Security Notices or NVD enrichment supply a description, affected version range, and CVSS vector.
Use-after-free memory corruption in Google Chrome before 154.0.8037.57 allows a remote, unauthenticated attacker to potentially escape the browser's renderer sandbox by persuading a victim to visit attacker-controlled web content. Exploitation requires mandatory user interaction (UI:R) and a high-complexity precondition such as precise heap grooming or a timing race (AC:H), and no public exploit has been identified at time of analysis. Google has issued a vendor-released patch in version 154.0.8037.57 that removes the vulnerable code path.
Out-of-bounds write (CWE-787) in Google Chrome permits a remote, unauthenticated attacker to corrupt memory in the browser when a victim opens or is redirected to attacker-controlled web content (UI:R), with high confidentiality and integrity impact and a scope change (S:C) that the independent assessment reads as the corruption crossing the renderer sandbox boundary. Exploitation is not turnkey: the high attack complexity (AC:H) reflects a precondition the attacker must satisfy - typically heap grooming, a race, or a specific object/memory state - before the write is reachable, which constrains reliability and mass exploitation. No public exploit identified at time of analysis, and the fixed build 154.0.8037.57 is already shipping, so exposure is limited to un-updated Chrome clients, making this a genuine high-priority browser issue rather than a top-tier emergency.
Risk cannot be meaningfully assessed from the supplied data: CVE-2026-95283 is recorded only as a vendor-reported entry from Ubuntu, with the description marked '[Unknown description]' and no CVSS score, CVSS vector, CWE, CPE, or affected source package provided. Because the impacted component, deployment mode, and attack surface are all unidentified, no statement can be made about what an attacker could achieve, which versions are affected, or what preconditions (authentication, user interaction, configuration) would be required. There is no KEV listing and no public exploit code identified at time of analysis, so the entry should be treated as an unassessed tracking record rather than a confirmed exploitable issue.
Race condition in Google Chrome's stable channel allows a remote, unauthenticated attacker who persuades a user to interact with crafted content to win a time-of-check/time-of-use window and affect data beyond the vulnerable component's security scope. The CVSS 3.1 base score of 8.0 (AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:N) reflects high confidentiality and integrity impact with no availability impact, tempered by high attack complexity; no public exploit code or CISA KEV entry was identified at time of analysis. The flaw is addressed in the 154.0.8037.57 stable channel update, and while the vendor feed tags it Information Disclosure, the scoring indicates a broader boundary-crossing defect, so this is best treated as a moderately high-priority browser issue with partly inferred impact.
No technical details are available for CVE-2026-95338: the description field is empty and no CVSS score, CVSS vector, CWE classification, CPE string, or affected package name was supplied alongside the record. The only identifying signal is that the issue was reported under the vendor:ubuntu source, indicating it is tracked as a vulnerability in a package distributed through Ubuntu's repositories and handled via the Ubuntu Security Notice process. Risk cannot be meaningfully assessed from the available data, and there is no public exploit identified at time of analysis; exploitation prerequisites, impact, and affected versions must be obtained from the Ubuntu Security Notice for this CVE before any triage decision is made.
Improper input validation in Google Chrome before 154.0.8037.57 allows unauthenticated remote attackers to achieve high confidentiality, integrity, and availability impact with a scope change, though exploitation requires the victim to visit attacker-controlled web content and overcome high attack complexity. The flaw, tracked as CVE-2026-95341 with a CVSS 8.3 base score, affects default installations of Chrome on all platforms and could permit renderer-level code execution or sandbox escape; no public exploit identified at time of analysis. Vendor-released patch 154.0.8037.57 addresses the issue.
Type confusion in Google Chrome builds prior to 154.0.8037.57 lets a remote attacker who convinces a victim to load attacker-controlled web content corrupt memory inside the renderer and potentially achieve code execution in that process. The flaw is rated CVSS 8.8 (AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H, CWE-843), meaning no authentication or privileges are required but exploitation is not zero-click: the victim must render a malicious page, and because scope is unchanged the initial compromise is confined to the renderer sandbox unless chained with a separate sandbox escape. No public exploit code has been identified at time of analysis and the issue is not listed as confirmed actively exploited (CISA KEV); a vendor-released patch is available in Chrome 154.0.8037.57.
CVE-2026-95277 is currently an effectively empty record: no description, CVSS score, CVSS vector, CWE mapping, affected product, version range, or CPE data has been published, and the only intelligence signal is attribution to the Ubuntu vendor reporting channel. Because the vulnerability class, attack vector, authentication requirements, and impacted component are all unknown, risk cannot be meaningfully assessed from the available data and no exploit status can be confirmed - there is no public exploit identified at time of analysis, and the entry is not listed as confirmed actively exploited (CISA KEV). Organizations should treat this identifier as a placeholder to be re-triaged once Ubuntu publishes a corresponding security notice (USN) or the upstream CVE record is populated.
An out-of-bounds memory write in Google Chrome allows an unauthenticated, remote attacker to corrupt memory and potentially execute arbitrary code or crash the browser when a victim opens a crafted web page, as reflected by the CVSS 3.1 vector AV:N/AC:L/PR:N/UI:R/S:C/C:H/I:H/A:H (9.6, scope-changed). The flaw is classified as CWE-787 (out-of-bounds write) and is addressed by the vendor's stable channel update for desktop, with the fix credits originating from the Ubuntu vendor reporting channel, meaning downstream Chromium packages are also in scope. At time of analysis there is no public exploit identified, no CISA KEV listing, and no EPSS data provided; however, because browser memory-corruption bugs of this class are historically attractive to exploit developers, this remains a high-priority patching item despite the absence of exploitation telemetry.
Improper input validation in Google Chrome before 154.0.8037.57 can be triggered when a victim loads attacker-controlled web content, yielding a scope-changed impact rated high for confidentiality, integrity, and availability (CVSS 8.3). Exploitation is network-reachable without credentials (PR:N) but demands victim interaction (UI:R) plus non-trivial preconditions such as specific memory state, heap grooming, or timing windows (AC:H), and the affected component is undisclosed because chromium issue 517730821 is access-restricted. No public exploit code and no confirmed active exploitation were identified at time of analysis; a vendor patch is available in the Chrome stable channel at 154.0.8037.57.
Use-after-free in Google Chrome before 154.0.8037.57 can be triggered when a victim visits attacker-controlled web content, potentially allowing a remote unauthenticated attacker to corrupt memory and escape the renderer sandbox to execute code or disclose information. The vendor-assigned CVSS 3.1 score is 8.3 (AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:H), reflecting mandatory user interaction and high attack complexity. Vendor-released patch: 154.0.8037.57; no public exploit was identified at time of analysis.
Use-after-free memory corruption in Google Chrome desktop builds prior to 154.0.8037.57 allows remote attackers to achieve high-integrity and high-confidentiality impact against a victim who loads attacker-controlled web content and performs some interaction with it, but exploitation is not trivially repeatable: with AC:H it depends on winning a race or reliably grooming heap layout rather than a single deterministic trigger. The flaw is reachable without authentication (PR:N) and the vector's scope change (S:C) indicates the assessed impact crosses a component boundary, yet no public exploit code has been identified at time of analysis and there is no CISA KEV listing, so this should be treated as a high-priority patching item rather than a confirmed in-the-wild threat. A vendor patch is available.
A race condition in the btrfs filesystem's relocation code (merge_reloc_roots() in fs/btrfs/relocation.c) allows a local user with an active balance or device add/remove/replace operation in progress to observe an inconsistent root->reloc_root pointer because the code clears that pointer without the memory barrier used elsewhere in the same code path. The affected code also clears the BTRFS_ROOT_DEAD_RELOC_TREE flag even when the root no longer points at the reloc root being cleaned up, so cleanup can target the wrong reloc root. The upstream fix routes cleanup through clear_reloc_root(), which issues the barrier after nulling ->reloc_root and before clearing the flag. Our independent assessment rates this far below the vendor-published CVSS 7.8: it is a local, high-complexity race that requires btrfs to be in active use and block-group relocation to be underway, is only meaningfully exploitable on weakly-ordered CPU architectures, and yields at most low integrity and availability impact with no confidentiality loss; there is no remote vector and no public exploit identified at time of analysis, and no KEV listing.
The Linux kernel's iwlwifi PCIe transport driver can leave dangling RX pointers (rx_pool, global_table, rxq and alloc_page) on the live transport object after iwl_pcie_rx_free() runs while unwinding a failed iwl_pcie_tx_init(), so a subsequent teardown or driver re-init/retry can touch already-freed RX state. Only systems with Intel WiFi hardware running kernels before 6.12.111, 6.18.53 or 7.2 are affected, and the trigger is an internal nic-init failure path rather than a directly attacker-controlled input, meaning exploitation is local with high complexity and existing low privileges (AV:L/AC:H/PR:L per the assessed vector). There is no public exploit identified at time of analysis and no KEV confirmation of active exploitation, and while the vendor/NVD score is 7.0 (High), the independent assessment rates this at roughly 3.6 (Low) because only integrity and availability of the device's own error/cleanup path are at stake.
Out-of-bounds reads and subsequent kernel instability in the Linux kernel's iwlwifi MVM driver (Intel WiFi) can be triggered when the WiFi firmware returns a malformed TX_CMD response whose frame_count does not match the actual response buffer layout. The affected code parses status entries by walking frame_count and then reads a trailing SCD SSN field without first validating that the buffer is long enough, so a malicious, compromised, or substituted device firmware can push the parser past the end of the payload, yielding limited information disclosure and, more likely, a kernel fault that takes down the driver or system. Systems running any Linux kernel built from the affected ranges (fixed in 6.12.111, 6.18.53, and 7.2) with an Intel WiFi adapter are exposed; there is no public exploit code and no confirmed active exploitation (CISA KEV) at time of analysis, and the practical attack surface is a firmware-to-host trust boundary rather than over-the-air radio traffic. Note that the supplied CVSS of 8.8 (AV:A/AC:L) overstates real-world risk; an independent assessment scores this AV:L/AC:H with confidentiality Low and availability High.
Out-of-bounds tid_data array access in the Linux kernel's iwlwifi mvm driver (CVE-2026-93790) is a theoretical flaw that writes the lq_color value into an unrelated tid_data slot, but it is not practically triggerable because multi-TID blockack is not in use, meaning the TFD loop counter 'i' is always 0 and the write remains in-bounds on unpatched systems. The affected component is the Intel iwlwifi mvm Wi-Fi driver in Linux kernel versions prior to 6.18.53, 6.12.111, and 7.2. An attacker within Wi-Fi radio range (adjacent, unauthenticated per the assessed CVSS vector) would need a contrived multi-TID blockack scenario to cause minor rate-control state corruption; no public exploit code exists and there is no evidence of active exploitation, so real-world risk is low.
An out-of-bounds slab read in the Linux kernel's SMB/CIFS client allows an attacker-controlled SMB1 server to make a victim's kernel read past the end of a kmalloc'd response buffer when that victim lists a directory on a mounted share. cifs_filldir() trusts the server-supplied de.namelen field from a TRANS2_FIND_FIRST/FIND_NEXT response without verifying that the name stays inside the response buffer, so an oversized FileNameLength drives memchr() (via filldir64()/verify_dirent_name()) beyond the allocation and produces a KASAN slab-out-of-bounds report; CVSS 3.1 is 8.1 (AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:N/A:H), with high confidentiality and availability impact from the out-of-bounds read. Exploitation requires the victim to mount a share from a hostile server using the deprecated SMB1 dialect and then browse it, so user interaction is required; no public exploit identified at time of analysis.
Permission widening in the Linux kernel's ksmbd in-kernel SMB3 server allows an authenticated SMB session holder to create files and directories whose inherited POSIX ACLs are recomputed incorrectly, discarding restrictive ACL_MASK entries that the parent directory's default ACL was intended to enforce, so newly created objects can end up more permissive than the parent's policy allows. The bug affects the ksmbd code shipped in Linux kernels before 6.18.53 and before 6.12.111, and in practice only matters where an SMB share exports a parent directory with a restrictive default POSIX ACL; userspace Samba deployments are unaffected. Vendor-released patches exist (6.18.53, 6.12.111 and 7.2) and no public exploit code has been identified at time of analysis.
A race condition in the Linux kernel's vhost-scsi backend allows a local, authenticated attacker with access to the vhost-scsi device to write command responses to an unrelated userspace memory region by replacing the memory table (VHOST_SET_MEM_TABLE) while SCSI commands are in flight. This can lead to host memory corruption and potential privilege escalation, though the attack requires a non-default vhost-scsi offload configuration and a narrow race window that makes reliable exploitation difficult. No public exploit has been identified at time of analysis; patches are available in Linux 6.12.111, 6.18.53, and 7.2.
A use-after-free in the Linux kernel's netfilter nfnetlink_log (NFLOG) subsystem lets a local attacker who can create and tear down network namespaces race packet logging against per-net state teardown, so that ops_free_list() frees the NFLOG pernet data while nf_log_packet() may still dereference the stale logger pointer under rcu_read_lock(). Impact is confined to local hosts running an affected kernel where the NFLOG logging path is in use and unprivileged network-namespace creation (or CAP_NET_ADMIN) is available; the flaw is a timing race, not a remotely reachable bug, and the assessed vector is AV:L/AC:H/PR:L. No public exploit code has been identified at time of analysis and the issue is not confirmed actively exploited (CISA KEV); vendor patches are available in Linux 6.12.111, 6.18.53, and 7.2.
Oversized SMBus block transfers in the Linux kernel I2C subsystem can reach tracepoints and native driver callbacks without ever being validated against I2C_SMBUS_BLOCK_MAX, allowing local memory corruption and out-of-bounds reads. The flaw is fixed by adding an early length check in __i2c_smbus_xfer(), which previously deferred validation to i2c_smbus_xfer_emulated() - too late for the smbus_write tracepoint and skipped entirely by adapters that supply a native .smbus_xfer implementation. The independently assessed vector (AV:L/AC:H/PR:L/UI:N/S:U/C:H/I:N/A:H) reflects that exploitation requires local access to an /dev/i2c-* interface plus a non-default configuration - SMBus tracing enabled (Bug 1) or the i2c-stub testing module loaded (Bug 2) - and default production kernels without tracing, i2c-stub, or unprivileged i2cdev access are not readily exploitable; no public exploit code was identified at time of analysis.
Local authenticated users on systems running Linux kernel 6.17 or newer with GPU drivers that exercise drm/pagemap device-memory migration can, on the rare migration-error path, cause DMA mappings to remain after pages are released because drm_pagemap_migrate_unmap_pages() is called after the pages array has been cleared. This can lead to limited integrity and availability impacts within the affected subsystem. The input CVSS of 8.8 substantially overstates the real-world risk; independent assessment rates it 5.3 (AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:L). No public exploit identified at time of analysis.
Out-of-bounds reads in the Linux kernel's Greybus audio topology parser allow a malicious or compromised Greybus module to leak kernel memory and crash the system. The module declares a small topology blob size but supplies large le32 section sizes (size_dais, size_controls, size_widgets) that are added to build the control, widget and route section offsets without ever being checked against the fetched blob, so gbaudio_tplg_parse_data() walks past the allocation. Exploitation requires the kernel to be built with Greybus audio support and a rogue module attached to the UniPro/Greybus interconnect - a rare hardware configuration - and no public exploit code has been identified at time of analysis. Note that while the supplied NVD-style vector rates this 8.8 with C:H/I:H/A:H under AV:A, our assessment is that the weakness is fundamentally a read-side bug requiring physical/hardware access (AV:P, C:H/I:N/A:H), and the 8.8 score overstates both the access path and the integrity impact.
Local privilege escalation-class memory-safety flaw in the Linux kernel's RDMA/bnxt_re driver allows an authenticated local user with access to the RDMA userspace-verbs interface (/dev/infiniband/uverbsN) to trigger use-after-free or double-free conditions by forcing udata copy failures after resource teardown, causing destroy callbacks to run again against freed driver state; a create-AH path can also leak a hardware object. The CVSS 3.1 vector is AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H (7.8), and exploitation requires a Broadcom NetXtreme-E adapter running bnxt_re in RoCE/RDMA mode, which is not a default configuration on general-purpose hosts. No public exploit was identified at time of analysis.
Out-of-bounds memory access in the Linux kernel PCI/pwrctrl driver for the Toshiba TC9563 PCIe switch, caused by the driver parsing both physical functions of the DSP3-integrated Ethernet MAC Endpoint when only one entry exists in the tc9563_pwrctrl->cfg array. Affected systems are those running Linux 6.19 or kernels built from commit 4c9c7be47310c1dbd7b6d37d45986123f5b133b4 up to the fixes in 7.2.6 / 7.3-rc1, but only when the specific TC9563 hardware is physically present, since the buggy code runs during driver probe; there is no network path or attacker-supplied runtime input. No public exploit has been identified at time of analysis (no CISA KEV listing, no POC, no EPSS signal provided), and while the published vector is 7.7, independent assessment rates real-world risk LOW with a modified vector of AV:L/AC:H/PR:H/UI:N/C:N/I:L/A:H, reflecting the hardware prerequisite as the dominant limiting factor.
A use-after-free in the Linux kernel's MD RAID5 Partial Parity Log (PPL) implementation lets a local attacker holding low privileges on a host that runs a PPL-enabled RAID5 array corrupt kernel memory, with the assessed impact being full loss of confidentiality, integrity and availability (CVSS 7.8, AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H) and a plausible path to kernel code execution or privilege escalation to root. The bug lives in ppl_do_flush(), whose loop keeps iterating over io->pending_flushes after ppl_io_unit_finished() has already freed the io unit; the fix is a single break statement that terminates the loop once the unit is released. Vendor patches are available across the affected stable branches, and no public exploit code was identified at time of analysis; exploitation is gated on reaching the PPL flush path at all (RAID5 with PPL journaling) and on winning the use-after-free window, so practical priority is moderate rather than urgent.
A use-after-free in the Linux kernel's VXLAN multicast database (MDB) handling allows a local attacker who holds CAP_NET_ADMIN (including within a user namespace) to corrupt kernel memory by manipulating VXLAN MDB entries and then triggering a flush. The bug lives in vxlan_mdb_flush(), which iterates the MDB hash list with hlist_for_each_entry_safe() under the false assumption that only the current entry can be removed; flushing the remotes of a (*, G) entry actually cascades into removal of the associated (S, G) entries, one of which may be the saved next-pointer, producing a double hlist_del() that writes LIST_POISON1 into LIST_POISON2 and yields attacker-influenced memory corruption (KASAN reported wild-memory-access). Per the assessed vector (CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H, 7.8) the impact is high for confidentiality, integrity and availability, but exploitation requires local privileges, CAP_NET_ADMIN, and a non-default VXLAN MDB configuration; there is no remote or unauthenticated path. No public exploit code was identified at time of analysis and the issue is not confirmed actively exploited (not in CISA KEV); vendor-released patches exist for the affected stable branches.
Page-table corruption can be triggered on LoongArch Linux kernels when the architecture fails to define DIRECT_MAP_PHYSMEM_END, causing get_free_mem_region() and mhp_get_pluggable_range() to fall back to the generic mm.h constant (1ULL << MAX_PHYSMEM_BITS) - 1 instead of the CPU's actual vmemmap-covered range. On LoongArch CPUs whose probed physical address bits (cpu_pabits from CPUCFG1) are fewer than MAX_PHYSMEM_BITS - notably the Loongson-2K series - a ZONE_DEVICE region can be allocated outside the vmemmap window, and vmemmap_populate() then wraps the memmap range around into low memory, silently corrupting page tables. The bug is local (CVSS vector PR:L), requires specific hardware plus CONFIG_SPARSEMEM_VMEMMAP and the ZONE_DEVICE/device-memory or amdkfd code path, and our independent assessment rates it AC:H with only C:L (versus the auto-generated 7.8 score with AC:L/C:H); no public exploit code or CISA KEV entry was identified at time of analysis.
A missing read memory barrier in the Linux kernel's in-kernel NFS server (nfsd) netlink status-dump path lets a low-privileged local user (CVSS PR:L) on weakly-ordered CPUs such as ARM or POWER read RPC status fields that have been concurrently modified by nfsd_dispatch, because the hand-rolled seqcount retry in nfsd_nl_rpc_status_get_dumpit() never orders its field reads ahead of the second counter check. The result is information disclosure of inconsistent or torn status data (CVSS C:H/I:N/A:H per the supplied vector), not memory corruption or code execution, and the reordering does not manifest on strongly-ordered x86/x86_64 hardware, which eliminates practical exploitability there. No public exploit code has been identified at time of analysis and there is no KEV confirmation of active exploitation, though our independent assessment scores the issue lower (AV:L/AC:H/PR:L/UI:N/S:U/C:L/I:N/A:L) than the supplied 7.1 base score because of the architecture restriction, the local-only vector, and the narrow timing race required against live NFS traffic.
Linux kernel NFS/RDMA servers using svcrdma accept RPC-over-RDMA Write and Reply chunks whose segcount field is zero, because xdr_check_write_chunk() only rejects segcount values above rc_maxpages and xdr_inline_decode(stream, 0) succeeds without advancing the cursor, causing pcl_alloc_write() to link a struct svc_rdma_chunk with ch_segcount == 0 onto the rc_write_pcl or rc_reply_pcl parsed-chunk lists even though RFC 8166 assigns no meaning to a chunk describing no remote buffer. The vulnerable path is reachable only on NFS servers configured with the RDMA transport over InfiniBand, RoCE, or iWARP, so an attacker must already be an RDMA peer able to send Write or Reply chunks to the server, and the independent assessment rates the realistic impact as availability-only (CVSS 3.1 AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:N/A:H) rather than the 9.1 input score, because an earlier patch in the same series already made pcl_for_each_segment() safe for ch_segcount == 0. No public exploit code and no confirmed active exploitation are identified at time of analysis, and vendor-released patches are available in 6.12.111, 6.18.50, 7.2.4, and mainline 7.3-rc1.
The Linux kernel's fsl-imx8mq-usb PHY driver leaks the Type-C (TCA) switch reference when probe fails after imx95_usb_phy_get_tca() has already succeeded, because the only release path lived in the driver's .remove() callback, which does not execute on probe failure. Only systems running NXP i.MX8MQ/i.MX95 hardware with the Type-C switch path active are affected, and triggering the bug requires high local privileges to force repeated probe failures; a single probe failure produces only a one-time leak of limited availability impact, and no public exploit has been identified at time of analysis (no KEV confirmation). The headline CVSS 7.4 with PR:N and full C/I/A:H is inflated relative to the assessed CVSS:3.1/AV:L/AC:H/PR:H/UI:N/S:U/C:N/I:N/A:L, which reflects a privileged, hardware-dependent, low-impact resource leak rather than a remotely exploitable compromise.
Failed connection acceptance in the Linux kernel's svcrdma transport (server-side NFS-over-RDMA) can corrupt reference-count and XArray registration state, leaking transport memory and mismatched device references. The bug is reachable only on hosts running the kernel NFS server over an RDMA fabric (InfiniBand or RoCE) where svc_rdma_accept() takes its error path before rpcrdma_rn_register() succeeds, and no public exploit has been identified at time of analysis. The assessed impact is primarily availability (repeated failed accepts leak svcxprt_rdma and its net/ns_tracker, leading to memory exhaustion) with a limited integrity effect from an erroneous erase of another caller's slot-0 entry; the CVSS vector carries PR:N, so no authentication is required, but attacks must originate from the adjacent RDMA fabric.
Use-after-free in the Linux kernel's server-side SUNRPC RPCSEC_GSS credential decoder (svcauth_gss_decode_credbody) leaves the pooled per-request credential (svcdata->clcred in gss_svc_data) holding a dangling gc_ctx.data inline pointer into a released request XDR page while gc_ctx.len still carries a stale length from an earlier request, so a downstream length-driven consumer such as gss_svc_searchbyctx() walks freed memory. Exposure is limited to hosts running an NFS server with RPCSEC_GSS/Kerberos security enabled (exports mounted with sec=krb5, krb5i, or krb5p); systems using only AUTH_SYS or running no NFS server are not affected. A remote, unauthenticated attacker who can reach the GSS-enabled service and craft a credential body that fails XDR decode can trigger the fault, but reliability is limited because the dangling state depends on residue from a prior request and on the request pages being released, so denial of service with limited information disclosure is the realistic outcome rather than assured code execution; no public exploit identified at time of analysis, and vendor patches are available for the affected stable trees.
A stack-based buffer overflow in Altera's Trusted Firmware port for SoC FPGA devices (the SoCFPGA branch through socfpga_v2.14.0) lets a local attacker who already controls the non-secure rich OS - root or kernel on the HPS ARM cores - corrupt the secure monitor's stack via the SMC interface and potentially escalate across the trust boundary into the secure world. The assessed vector (CVSS:3.1/AV:L/AC:L/PR:H/UI:N/S:C/C:L/I:H/A:H) confirms this is an access-gated privilege/trust-boundary escalation requiring an existing high-privileged foothold, not a remotely reachable default-configuration flaw; impact to integrity and availability is high and scope is changed, but confidentiality impact is limited. No public exploit code has been identified at time of analysis and the issue is not confirmed as actively exploited (not in CISA KEV).
Heap over-read in DCMTK through 3.7.0 allows unauthenticated remote attackers to crash applications or leak adjacent heap memory by supplying a crafted DICOM instance whose declared NumberOfFrames exceeds the actual PixelData buffer length. The flaw resides in the ConcatenationLoader module (dcmfg) and is triggered when loading multi-frame DICOM concatenations; exploitation requires an ingestion path that processes untrusted instances (e.g., a DICOM network storage service), as offline viewers typically require a user to open the file. No public exploit code is identified at time of analysis, and a vendor patch is available via upstream commits.
Denial of service in NodeRedis redis-parser through 3.0.0 lets a hostile or hijacked Redis endpoint crash the Node.js client process by returning a crafted RESP multi-bulk header whose declared length exceeds 2^32-1, triggering an uncaught RangeError during parsing. The flaw is reachable by unauthenticated remote attackers (CVSS 4.0 AV:N/AC:L/AT:N/PR:N/UI:N) but its impact is strictly availability (VA:H) - no confidentiality, integrity, or scope impact - so a malicious server, a compromised legitimate Redis instance, or an on-path injector in the RESP response stream can only take down the parsing process. Rated 8.7 (HIGH) by CVSS 4.0, with no public exploit identified at time of analysis and no confirmed vendor-released patch, though the issue is publicly tracked upstream.
Command injection in rpm enables an attacker to run arbitrary shell commands as the invoking user when a victim installs, rebuilds, or otherwise processes a malicious source RPM (.src.rpm) that contains a %() macro construct in a source or spec file basename. The flaw is local and requires user interaction, with impact confined to the privileges of the user running rpm or rpmbuild, typically non-root; build farms and CI/CD pipelines that automatically handle untrusted SRPMs are the highest-risk targets. Red Hat CPEs list RHEL 6 through 10 and Red Hat Hardened Images, and no public exploit was identified in the provided intelligence; monitor the vendor advisory for exploitation and patch updates.
Local attackers can achieve arbitrary code execution by tricking a user or automation into processing a crafted manifest file with 'rpm -q -p', due to unexpected macro expansion of manifest entries in RPM before they are opened. This command injection flaw (CWE-78) affects Red Hat Enterprise Linux 6 through 10 and Red Hat Hardened Images, with a CVSS 3.1 base score of 7.8 (HIGH, AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H); exploitation requires user interaction but no privileges, and impact is limited to the privileges of the rpm process. No public exploit code or active exploitation has been identified at time of analysis.
Privilege escalation in Rapid7 Velociraptor allows an authenticated user holding only the 'investigator' role to execute arbitrary VQL statements with administrator privileges on the Velociraptor server by populating the internal 'compiled_collector_args' field when submitting a hunt-creation/scheduling API request. Because the server accepts this internal field from the caller, the attacker-supplied pre-compiled VQL is trusted and the ACL checks that normally gate artifact compilation are skipped (CVSS 9.9, AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H). Exploitation requires a valid investigator-or-higher account that can schedule hunts, making this an insider/privilege-escalation path rather than an unauthenticated remote compromise; no public exploit identified at time of analysis, and no confirmed actively exploited (CISA KEV) status was provided. The vendor advisory is published at https://docs.velociraptor.app/announcements/advisories/cve-2026-19072/.
An out-of-bounds write in GIMP's GIMPressionist plug-in allows an attacker to corrupt memory by convincing a user to load a malicious preset file, potentially leading to a crash or arbitrary code execution. The flaw is an unauthenticated local attack (CVSS PR:N) that requires user interaction (UI:R) to open a crafted preset file, and it affects GIMP as packaged in Red Hat Enterprise Linux 6 through 10 plus likely upstream GIMP versions. No public exploit has been identified at time of analysis, and the overall risk is assessed as genuine but not urgent due to the requirement for user interaction with a crafted file.
Heap-based buffer overflow in the Nanomsg WebSocket transport (versions 0.5-beta through 1.x before 1.2.3) allows remote attackers to corrupt heap memory by sending an oversized Sec-WebSocket-Version header during the WebSocket handshake, which is processed before any authentication or user interaction takes place. Exploitation requires that the target application actually enable and expose nanomsg's ws transport; TCP- or IPC-only deployments are unaffected. The flaw yields high integrity and availability impact, and while reliable code execution depends on heap layout and platform, no public exploit code has been identified at time of analysis and the vulnerability is not confirmed as actively exploited (not in CISA KEV).
Denial of service in the MuJS embeddable JavaScript interpreter (git revision e892c9fdb) allows remote attackers to crash the hosting process by supplying JavaScript that builds an excessively large numeric array index, which jsR_isindex() in jsrun.c converts from floating point to integer without validating the value fits in the target integer range. Per the assessed CVSS vector (AV:N/AC:L/PR:N/UI:N/S:U/C:N/I/A:H), the attack is network-reachable and unauthenticated, but the impact is availability-only: no confidentiality or integrity loss, with exploitability gated by whether the embedding application evaluates untrusted JavaScript. Exploitation requires the host to feed attacker-controlled script to the embedded interpreter (for example, opening a document or file that MuJS parses), and no public exploit code or CISA KEV entry was identified at time of analysis; EPSS is 0.15% (4th percentile), indicating low near-term exploitation likelihood. The upstream project has published a fix commit, but no tagged patched release has been independently confirmed.
Crafted DNS traffic parsed by the nDPI 5.1.0 deep-packet-inspection library can terminate the hosting process wherever that code runs on a strict-alignment CPU (ARM, MIPS, SPARC and similar) or in an UndefinedBehaviorSanitizer/fuzzing build, because the DNS dissector and the serializer deserialization path dereference byte buffers at arbitrary packet offsets through uint16_t/uint32_t/uint64_t pointers without any alignment check. The flaw is remotely reachable and unauthenticated (CVSS 3.1 AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H, 7.5), but impact is limited to availability - process termination, no confidentiality or integrity loss - and the decisive precondition is the build/CPU target: mainstream x86-64 release builds silently tolerate the misaligned load and generally will not crash. It is not listed in CISA KEV and no public exploit identified at time of analysis, although the upstream fix PR does ship fuzz corpus inputs (issue3213_trigger1-3) that reproduce the fault, and EPSS data was not available for this record.
Denial of service in hiredis post-v1.5.0 (commit 29ea279) occurs when the RESP aggregate parser performs an uncontrolled memory allocation based on an oversized element or length count supplied in a Redis server reply. A hiredis-based client that connects to a malicious or compromised Redis server, or whose unencrypted connection is subject to man-in-the-middle redirection, can be driven into memory exhaustion and crash or unresponsiveness; no authentication on the hiredis side and no user interaction are required. The provided data does not include a CISA KEV entry, and no public exploit identified at time of analysis, so exploitation should be assessed as requiring an attacker-controlled or untrusted Redis endpoint rather than as confirmed active exploitation.
Type confusion in the Perl DBI module's fetchrow_hashref implementation allows an application that assigns a non-string scalar (integer/IV or float/NV) to the FetchHashKeyName attribute and then calls fetchrow_hashref to trigger an out-of-bounds memory read, typically crashing the Perl process (denial of service) and, in principle, exposing adjacent heap memory in error output. Affected are DBI releases before 1.654 on any platform where the module is installed; the CVSS 3.1 vector (AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H, 7.5) rates this unauthenticated and remotely triggerable, though independent assessment raises attack complexity to AC:H because exploitation requires the hosting application to propagate attacker-controlled data into a statement-handle attribute that is normally set by the application itself. No public exploit code has been identified at time of analysis, no confirmed active exploitation (not in CISA KEV), and EPSS is low at 0.21% (10th percentile), reflecting the narrow precondition of application-specific attribute misuse.
Denial of service in NanoSVG at commit 239e102ec affects applications that parse attacker-supplied SVG documents, where a crafted arc command with extreme radius values drives the arc delta-angle calculation to NaN and the subsequent unchecked NaN-to-int conversion causes undefined behavior and process termination. The published CVSS vector is 7.5 (AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H), so the unauthenticated network attack path is constrained by the requirement that the target actually process a malicious SVG, and manual SVG opening can effectively require user interaction; EPSS is 0.16% and no public exploit identified at time of analysis.
Heap out-of-bounds write in QuickJS's bytecode deserializer (JS_ReadFunctionTag) lets a crafted serialized blob corrupt heap memory, giving an attacker the potential for arbitrary code execution in the context of the host process. The flaw is only reachable when the embedding application calls JS_ReadObject with the JS_READ_OBJ_BYTECODE flag on attacker-controlled input - not the default pattern of executing JavaScript source - so exposure is confined to plugin, cache, or IPC loaders that treat serialized bytecode as untrusted data; where that path is exposed, the flaw is remotely triggerable by an unauthenticated attacker (CVSS vector PR:N) with a single malformed blob and no user interaction. Public exploit code has not been identified at time of analysis (EPSS 0.17%, 5th percentile), and while the vendor score is 9.8 with AC:L, an independent assessment rates attack complexity as high (CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:H) because exploitation depends on a non-default deserialization path in the hosting application.
Denial of service in PHP: attacker-influenced input passed into a recursive routine can exhaust process resources and terminate the PHP worker or interpreter (CVSS 3.1 base 7.5, AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H), with impact confined to availability and no data exposure or code execution. The flaw is network-reachable, unauthenticated, and low-complexity per the vector, so any deployment that feeds untrusted remote input into the affected recursive code path is exposed, while deployments that only pass trusted input are not. No public exploit code or confirmed active exploitation was identified at time of analysis, and the upstream PHP security advisory GHSA-rgrp-mwpx-f6rm is the authoritative source for the exact affected and fixed versions.
Crafted MAT4 (MATLAB v4) files cause an integer overflow in libsndfile 1.2.2's mat4_read_header() parser, crashing the process and producing a denial of service. The flaw is network-reachable without authentication (CVSS 3.1 PR:N), but our assessment requires a victim to open the file or a backend service to ingest untrusted audio (UI:R); impact is limited to availability, with no code execution or data disclosure. No public exploit code and no vendor-released patch are confirmed at time of analysis.
Denial of service in NanoSVG (commit 239e102ec) where a crafted SVG document carrying oversized geometry coordinates drives the rasterizer's fixed-point-scaled edge values past the range of a 32-bit int during nsvg__addActive(), producing undefined behavior that can terminate the process. The issue is rated CVSS 3.1 7.5 (AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H), i.e. unauthenticated and low-complexity, but availability-only: there is no code execution, data disclosure, or integrity impact per the assessed vector. Exploitation requires the embedding application to actually rasterize attacker-supplied SVG content rather than merely parse it, so tools that only handle locally trusted assets face a local-input scenario; no public exploit code was identified at time of analysis and the flaw is not confirmed actively exploited (not in CISA KEV), though the vendor issue tracker reference (GitHub issue 292) makes root-cause details publicly discoverable.
TLS client-authentication bypass in the RabbitMQ trust-store plugin lets a remote unauthenticated attacker whose forged self-signed certificate carries the issuer DN and serial number of any whitelisted certificate pass mutual-TLS authentication. Only deployments that have the rabbitmq_trust_store plugin enabled and wired in as the TLS verify_fun for a client-certificate listener are affected, and an attacker must first learn or guess at least one whitelisted certificate's {Issuer, Serial} pair, which is non-secret and is exposed by the plugin's list command, logs, or any copy of the certificate. All RabbitMQ server releases prior to 3.13.15, 4.0.20, 4.1.11, 4.2.6, and 4.3.0 are vulnerable; vendor-released patches exist in those versions, and no public exploit identified at time of analysis.
Memory-exhaustion denial of service in RabbitMQ's Web-MQTT plugin (rabbitmq_web_mqtt) lets an unauthenticated remote attacker crash a broker node by sending a single WebSocket frame that is highly compressed and inflates to gigabytes in memory. Affected are RabbitMQ server releases prior to 3.13.15, 4.0.20, 4.1.11, 4.2.6 and 4.3.0 whenever the Web-MQTT plugin is enabled and its listener (TCP 15675/15676) is network-reachable; the attacker only needs to negotiate RFC 7692 permessage-deflate during the upgrade and then send a zlib bomb, since decompression happens in the connection process before any MQTT CONNECT credentials are evaluated. No public exploit code was identified at time of analysis, exploitation requires the non-default plugin to be enabled, and the impact is availability-only (C:N/I:N/A:H) against the affected node. Vendor-released fixes are available in 3.13.15, 4.0.20, 4.1.11, 4.2.6 and 4.3.0.
Denial of service in RabbitMQ Server (versions prior to 4.3.0, 4.2.6, 4.1.11, 4.0.20, and 3.13.15) allows an authenticated AMQP 0-9-1 client with publish permission to exhaust broker memory by declaring an excessive content-header BodySize and streaming message fragments without completing assembly. The unvalidated uint64 BodySize bypasses the max_message_size check, which only runs at assembly completion, causing unbounded memory accumulation in the reader process until the memory alarm triggers or the node runs out of memory, degrading all publishers cluster-wide. No public exploit code has been identified at time of analysis, and vendor patches are available.
Denial of service in RabbitMQ versions prior to 4.2.7 and 4.3.1 allows any authenticated AMQP 0-9-1 client to crash the entire Erlang VM, taking down all vhosts and connections, by exhausting the atom table with roughly one million requests. The attack exploits the direct reply-to feature: queue names or routing keys beginning with 'amq.rabbitmq.reply-to.' cause an ETF-encoded pid node name to be passed to binary_to_atom without using binary_to_existing_atom, and because atoms are never garbage collected, the try/catch cannot reclaim them. Exploitation requires a valid AMQP connection (any credential, any permission level) and the ability to send the large number of operations; no public exploit identified at time of analysis.
Unauthenticated remote attackers can crash any RabbitMQ node with the AMQP 1.0 listener enabled by sending a single ~19-byte frame to TCP port 5672, causing the Erlang VM to exhaust heap memory and terminate, which takes down all tenants and protocols on that node. The flaw is an unbounded-allocation bug (CWE-770) in the AMQP 1.0 framing code that is reached before authentication completes, so no credentials or user interaction are required; only network reachability to the broker's AMQP 1.0 port is needed. Affected releases are RabbitMQ prior to 3.13.15, 4.0.20, 4.1.11, and 4.2.6, which contain the vendor-released fix; no public exploit code has been identified at time of analysis, and the issue is scored high severity (CVSS 8.2 v4.0 / assessed CVSS 3.1 7.5) with availability-only impact.
The RabbitMQ management plugin (rabbit_mgmt_cors.erl) reflects an attacker-controlled Origin header straight back in Access-Control-Allow-Origin while also returning Access-Control-Allow-Credentials: true, so when an operator has configured cors_allow_origins = "*" a hostile web page can make a signed-in administrator's browser issue authenticated, state-changing requests to the management HTTP API. Exploitation is conditional rather than automatic: it requires the non-default wildcard CORS setting, a live cached HTTP Basic-auth administrator session, and the administrator being lured to the attacker's page (user interaction), and no public exploit code has been identified at time of analysis; assessed severity is CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:N. The issue affects the management plugin in RabbitMQ releases prior to 3.13.17, 4.0.22, 4.1.13 and 4.2.6.
RabbitMQ brokers that use the OAuth2 plugin can silently fall back to disabled TLS certificate verification when no CA bundle can be located, allowing an attacker already positioned as a network man-in-the-middle to forge the JWKS response and inject arbitrary JWTs that the broker will accept as authentic. The flaw affects RabbitMQ server versions prior to 3.13.15, 4.0.20, 4.1.11, 4.2.6, and 4.3.0, and exploitation requires four conditions to align: the OAuth2 plugin must be in use, no cacertfile may be configured, the host or container's OS CA bundle must be empty or unreadable, and the attacker must hold a MitM position on the path to the identity provider's JWKS endpoint. Unauthenticated remote attackers are the relevant threat actor per the CVSS vector (PR:N), but the precondition stack and the required network position make this a heavily gated authentication-bypass rather than a blanket internet-exploitable critical; no public exploit code and no confirmed active exploitation were identified at time of analysis.
Stored cross-site scripting in the RabbitMQ management UI allows an attacker who can obtain a TLS client certificate signed by a CA the broker trusts to embed JavaScript in the certificate's Subject DN, which is rendered unescaped in the connection detail page and executes in an administrator's browser session when that connection is viewed. Success enables full administrative account takeover within the management UI - creating users, exporting definitions, and performing any action the viewing admin can perform. Exploitation is heavily gated: the TLS listener must be configured with ssl_options.verify = verify_peer (not the default), the certificate must be signed by a CA in the broker's trust store (self-signed certificates do not work), and an administrator must open the affected connection page; the management UI's CSP includes 'unsafe-inline', so inline script is not blocked. Patched versions are 3.13.15, 4.0.20, 4.1.11, and 4.2.6; no public exploit code and no confirmed active exploitation were identified at time of analysis.
Use-after-free in librsvg during rendering of crafted SVG documents with nested XInclude directives and duplicate entity declarations allows unauthenticated local attackers to crash or potentially execute code in applications that use the library. Tracked as CVE-2026-96889 and reported by Red Hat, the flaw affects librsvg as shipped in Red Hat Enterprise Linux 6, 7, 8, 9, and 10. Exploitation requires user interaction (UI:R) because a victim must open, preview, or thumbnail a malicious SVG; the CVSS 3.1 score is 7.8. No public exploit code was identified at time of analysis, and independent assessment indicates that arbitrary code execution is unproven-only denial of service or crash is reliably reachable.
Deeply nested JSON arrays and objects drive JVM heap exhaustion in Typelevel Jawn versions before 1.7.0, where the thrown java.lang.OutOfMemoryError is a fatal Scala/JVM error that neither scala.util.Try nor cats.effect.IO routinely intercepts, turning untrusted JSON into a denial-of-service condition. The assessed impact is availability-only (CVSS:3.1 7.5, AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H), so remote unauthenticated attackers need only reach an endpoint that feeds attacker-controlled input into a pre-1.7.0 Jawn parse call, directly or transitively through circe; services that only parse trusted or internal JSON are not exposed. No public exploit code has been identified at time of analysis and the issue is not listed in CISA KEV, but this is a genuine, middle-tier priority for any internet-facing JSON API on an affected version; fixed in 1.7.0, which introduces a configurable depth cap.
Denial of service in the Jawn JSON parser for Scala (all artifacts prior to 1.7.0) lets a remote, unauthenticated attacker exhaust CPU by feeding JSON through the streaming AsyncParser in many tiny chunks, forcing repeated rescans of a single incomplete token. Exploitation is bounded by two preconditions: the target must parse attacker-influenced input incrementally via AsyncParser rather than parsing a fully buffered document, and the attacker must be able to influence chunk boundaries so that one token is split across many small reads - hence the assessed CVSS 3.1 vector AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:N/A:H, an availability-only (HIGH) impact with no confidentiality or integrity effect. No public exploit code and no confirmed active exploitation were identified at time of analysis; the vendor fixed the issue in version 1.7.0.
Denial of service in JLine's built-in less viewer affects all 3.x releases from 3.0.0 before 3.30.15 and 4.x releases before 4.3.1: a user-supplied search or display-filter pattern containing nested quantifiers is handed straight to Java's backtracking regex engine and applied repeatedly across file content, so a pattern run against non-matching lines can consume excessive CPU and block the session thread indefinitely. The practical risk profile is multi-user Telnet or SSH deployments that embed the less viewer, where repeated sessions can exhaust a bounded worker pool and deny service to other users; the independent assessment scores this as availability-only (CVSS 3.1 A:H, no confidentiality or integrity impact) and notes that an interactive session foothold and viewable file content are prerequisites, which the published vector's PR:N rating does not reflect. No public exploit code and no confirmed active exploitation were identified at time of analysis, so exploitation is opportunistic rather than observed. The fix is the SafeRegex wrapper plus regex timeouts shipped in 3.30.15 and 4.3.1.
Denial of service in JLine3's built-in grep command (org.jline:jline-builtins) lets anyone who can supply a search pattern hang a worker thread indefinitely via catastrophic regular-expression backtracking. The builtin - reachable in applications that embed JLine3 and expose the shell to remote or interactive users - rewrites non-line-regexp patterns as .*pattern.* before compiling them with Java's backtracking regex engine, so a short input such as (a+)+b tested against non-matching text causes exponential CPU consumption (CWE-1333; CVSS 3.1 base 7.5 upstream, while the assessed vector requires low privileges, PR:L). Impact is availability-only (C:N/I:N/A:H) with no confidentiality or integrity loss; publicly available exploit code exists in the form of the advisory's two-line PoC, but the issue is not listed in CISA KEV and no confirmed active exploitation has been reported. Vendor-fixed releases are available: 3.30.15 on the 3.x line and 4.3.1 on the 4.x line.
Local privilege escalation to root in Flatpak before 1.18.1 lets an unprivileged user holding an active local session abuse the revokefs writer used by flatpak-system-helper: by opening two concurrent revokefs sessions and planting a symlink from one session directory into the other, the attacker retains a file descriptor into a peer session and keeps modifying its files after the system helper has already validated and imported them. The practical payoff is tampering with ostree commit objects in the system repository after signature verification, which translates into root-controlled writes to attacker-chosen paths. Exploitation requires local access and winning a TOCTOU race against the dual-session setup (CVSS AV:L/AC:H/PR:N), and no public exploit code has been identified at time of analysis; the issue is not confirmed as actively exploited (not in CISA KEV).
A heap buffer overflow in BusyBox's romfs filesystem image parsing path can be triggered when the utility processes an attacker-crafted romfs image whose volume ID metadata is read with an unbounded strlen, producing an out-of-bounds heap write (CWE-787) rated CVSS 3.1 base 7.3. Because the attack vector is local (AV:L), the malicious image must be delivered to the target system - for example via removable media, a downloaded archive, a firmware/container image, or a package - and some process or user must then parse or mount it (UI:R); it is not remotely triggerable over a network. Under the assessed vector (AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H) no privileges are required, and successful corruption could yield code execution in the context of the parsing process with high impact to confidentiality, integrity, and availability; no public exploit or POC was identified in the supplied intelligence and there is no confirmed active exploitation.
Pre-authentication heap memory corruption in BusyBox's built-in TLS implementation can be triggered by a remote, unauthenticated attacker who sends a single crafted ClientKeyExchange handshake message to a BusyBox process acting as a TLS server, corrupting the heap and crashing the process (CVSS 3.1 base 7.5, availability-only with no confidentiality or integrity impact rated). Exploitation requires a BusyBox binary compiled with TLS support and reachable in a server role, a deployment pattern that is the exception rather than the norm since BusyBox TLS is most often used outbound (for example by wget https), which sharply narrows the realistically exposed population. No public exploit code or confirmed active exploitation (CISA KEV) was identified in the available intelligence; the issue is tracked by Red Hat via its CVE page and Bugzilla, with no public exploit identified at time of analysis.
Path traversal in Flatpak's `flatpak build-init` command allows an attacker-controlled SDK container to write arbitrary files outside the developer's working directory when that developer builds against the SDK using both the `--writable-sdk` and `--sdk-extension` flags. The root cause (CWE-22) is that the SDK's declared extension-point `directory` value is resolved through a path helper that does not reject `..` components, so a malicious SDK can steer writes outside the intended build root. The independently assessed vector (CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H) is a local, user-interaction-dependent action rather than the remote 9.8 originally submitted; no public exploit code or confirmed active exploitation has been identified at time of analysis, and the attacker must convince the developer to consume a hostile SDK, which severely limits real-world exposure.
Message-filtering bypass in xdg-dbus-proxy before 0.1.9 lets a malicious or already-compromised sandboxed application - most notably a Flatpak app, but also deployments under frameworks such as Firejail - escape its confinement and execute arbitrary code on the host by attaching a forged reply serial number to non-reply D-Bus messages. The flaw carries a CVSS 4.0 base score of 8.7 (AV:N/AC:L/PR:N/UI:P, with high confidentiality, integrity and availability impact scoped to the sandbox-escape boundary) and is rated a genuine high-priority sandbox-escape issue rather than a CVSS-inflated paper risk; the practical precondition is that an attacker must first get code running inside a sandbox that routes its session bus through the vulnerable proxy, typically by tricking a user into installing or launching a hostile or trojanized app. No confirmed active exploitation (CISA KEV) and no public exploit code were identified at time of analysis; vendor-released patch: 0.1.9.
Apache Tomcat Native 1.3.0 through 1.3.8 and 2.0.0 through 2.0.15 can, under a narrow timing window, down-grade the client-certificate verification requirement for connections, weakening mutual-TLS authentication for deployments that enforce client-cert auth through the APR/OpenSSL-based connector. The flaw is a CWE-366 race condition within a thread (CVSS 3.1 base 7.4, AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:N), so an unauthenticated network attacker must win a non-deterministic race to succeed, making exploitation probabilistic rather than guaranteed. Only Tomcat instances that actually use Tomcat Native and are configured for client-certificate verification are exposed; no public exploit identified at time of analysis, and no CISA KEV confirmation of active exploitation. Vendor-released fixes are available in 2.0.16 and 1.3.9.
Insecure-by-default OpenSSL options in Apache Tomcat Native (versions 2.0.0 through 2.0.15 and 1.3.0 through 1.3.8) silently weaken the TLS layer used by the APR/OpenSSL connector, enabling insecure client-initiated renegotiation, disabling the Extended Master Secret extension, ignoring unexpected EOF (truncation), and permitting key exchanges without (EC)DHE forward secrecy. The vendor scored this 9.1 (CVSS:3.1/AV:N/AC:L/PR:N/UI:N/C:H/I:H/A:N), but our independent assessment rates it substantially lower - CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:L/A:N - because meaningful abuse requires an on-path/man-in-the-middle attacker who can actively position themselves between client and server, and because the four options are only applied when the deployment has not explicitly overridden them. Only Tomcat instances configured with the Apache Tomcat Native (APR) TLS connector at an affected version are exposed; JSSE/NIO-based connectors are unaffected, and no public exploit code or CISA KEV entry was identified at time of analysis. Vendors/users should upgrade to Tomcat Native 2.0.16 or 1.3.9.
Denial of service via JVM crash affects Apache Tomcat Native versions 2.0.0 through 2.0.15 and 1.3.0 through 1.3.8, where a buffer over-read (CWE-126) during the TLS handshake lets a malicious client terminate the server process, rated CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H. The flaw is reachable remotely and unauthenticated with no user interaction, but only where the APR/OpenSSL-based native TLS connector is actually deployed and handling TLS connections; Tomcat instances using the pure-Java JSSE/NIO connector are not affected. Impact is confined to availability - no confidentiality or integrity loss, no data exposure and no code execution - and no public exploit was identified at time of analysis, with no CISA KEV entry.
WebSocket message smuggling in Apache Tomcat allows a remote, unauthenticated attacker to inject or reorder messages within a compressed WebSocket stream, corrupting application-level message boundaries and undermining integrity of data exchanged over the connection. The flaw affects the 11.0.0-M1 through 11.0.25, 10.1.0-M1 through 10.1.59, and 9.0.0.M1 through 9.0.121 release lines, plus the end-of-support 8.5.0-8.5.100 and 7.0.56-7.0.109 branches, but it only triggers when the per-message-deflate compression extension has been negotiated on the WebSocket connection; deployments that serve only standard HTTP or that run WebSocket endpoints without compression are not exposed. There is no confirmed active exploitation (CISA KEV does not list this issue), no public exploit code identified at time of analysis, and CISA's SSVC assessment rates exploitation as 'none', making this a genuine but conditionally-gated integrity flaw rather than a drop-everything emergency.
HTTP/2 request smuggling in Apache Tomcat can mix up request headers between concurrent streams, allowing unauthenticated remote attackers to potentially read or alter other users' request data when the connector has HTTP/2 explicitly enabled. The flaw is a regression from the fix for CVE-2026-41293 and affects only Tomcat 11.0.22 through 11.0.25, 10.1.55 through 10.1.59, and 9.0.118 through 9.0.121; exploitation is rated high complexity (AC:H) with limited confidentiality and integrity impact (C:L/I:L). No public exploit code has been identified at time of analysis, and the independent assessment rates this as a moderate-priority, genuine protocol-layer flaw rather than a critical emergency.
Authentication can succeed when it should have been rejected in Apache Tomcat deployments that use CLIENT_CERT (mutual-TLS) authentication with soft fail disabled, allowing network attackers to reach resources that the mutual-TLS configuration was meant to protect. The affected range is broad - 9.0.92 through 9.0.121, 10.1.22 through 10.1.59, and 11.0.0-M14 through 11.0.25 - but the flaw is gated on two non-default conditions (the CLIENT_CERT authenticator in use and soft fail turned off) and, per the vendor, only manifests in some scenarios, so it is not universally reproducible even on correctly configured vulnerable instances. No public exploit code has been identified at time of analysis, and no evidence of active exploitation is present; the attacker needs network reachability to the mTLS-protected listener.
Denial of service in Apache Tomcat's asynchronous WebSocket write path allows remote, unauthenticated attackers to exhaust server resources by triggering lost write timeouts, so non-blocking WebSocket sends are never released after their effective lifetime. All maintained branches are affected - 11.0.0-M1 through 11.0.25, 10.1.0-M1 through 10.1.59, and 9.0.0.M1 through 9.0.121, plus the end-of-life 8.5.0-8.5.100 and 7.0.43-7.0.109 lines - although exposure is bounded to instances whose deployed applications actually perform asynchronous (non-blocking) WebSocket writes; servers with no WebSocket endpoints or only synchronous writes are not exposed. The flaw is rated CVSS 7.5 (AV:N/AC:L/PR:N/UI:N, availability-only) and CISA's SSVC assessment records no observed exploitation, with no public exploit code identified at time of analysis; the primary risk is sustained resource pressure that degrades or halts the affected connector rather than data compromise.
Apache Tomcat 11.0.19 through 11.0.25, 10.1.53 through 10.1.59, and 9.0.116 through 9.0.121 can cause an unrelated user's request to fail when a malformed request triggers incomplete cleanup of request-processing state on a reachable HTTP connector. No authentication or user interaction is required, but the effect is fundamentally gated by a race window - the advisory itself states the failure occurs only 'depending on timing' - so it is non-deterministic and cannot be reliably triggered on demand, and the realistic worst case is intermittent failure of another user's individual request rather than sustained denial of service, data disclosure, or code execution. No public exploit code and no confirmed active exploitation were identified at time of analysis; the 7.3 base score reflects low impact across confidentiality, integrity, and availability, though our assessment rates attack complexity as High rather than Low.
Denial of service in Apache Tomcat allows an attacker to pin an AJP processing thread through a single AJP request, exhausting the connector's limited thread pool (CWE-770, allocation of resources without limits or throttling). The affected range is broad: 11.0.0-M1 through 11.0.25, 10.1.0-M1 through 10.1.59, 9.0.0.M1 through 9.0.121, plus the end-of-life branches 8.5.0-8.5.100 and 7.0.0-7.0.109, which remain affected even though they are no longer supported. There is no public exploit identified at time of analysis, no KEV listing, and no EPSS score in the input; the independent assessment rates the issue CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H (availability-only impact), but exploitation requires the AJP connector to be enabled and network-reachable, which has not been the default configuration since Tomcat 9.0.31.
Uncontrolled resource consumption in Apache Tomcat's WebSocket close-message handling allows unauthenticated remote attackers to exhaust server resources and degrade availability on instances that expose WebSocket endpoints. Per the assessed vector (CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L), the impact is limited to availability (Low) with no confidentiality or integrity effect, and no user interaction is required. Affected releases are 11.0.0-M5 through 11.0.25, 10.1.8 through 10.1.59, 9.0.74 through 9.0.121, plus the end-of-life 8.5.88 through 8.5.100 line; fixes are available in 11.0.26, 10.1.60, and 9.0.122. There is no confirmed active exploitation (not in CISA KEV) and no public exploit identified at time of analysis, but the low attack complexity and unauthenticated network reachability make prompt patching advisable for any deployment using WebSockets.
Apache Tomcat's HTTP/2 connector harbors a race condition (CWE-362) that allows unauthenticated remote attackers to inject trailer fields into another user's concurrent HTTP/2 request, resulting in high confidentiality and integrity impact without affecting availability. Affected versions are 11.0.0-M1 through 11.0.25, 10.1.0-M1 through 10.1.59, 9.0.39 through 9.0.121, and the end-of-life 8.5.59 through 8.5.100; exploitation is conditional on winning a concurrency race and only applies when the HTTP/2 upgrade protocol is enabled and actively serving traffic. No public exploit code or active exploitation has been identified at time of analysis, and the overall threat is considered credible but conditional.
Authentication bypass in Apache Tomcat allows unauthenticated remote attackers to reach WebSocket endpoints that are supposed to be protected by container-managed security constraints, simply by referencing the endpoint under an alternate name so the constraint is never matched. Exploitation requires only that the application expose one or more WebSocket endpoints and rely on container-level (web.xml security-constraint) authentication or authorization rules to guard them; no credentials, no user interaction, and no non-default configuration beyond that are needed, and per the CVSS vector (PR:N) the attack is unauthenticated and network-reachable. Affected lines are Tomcat 11.0.0-M1 through 11.0.25, 10.1.0-M1 through 10.1.59, 9.0.0.M1 through 9.0.121, plus end-of-support 8.5.0-8.5.100 and 7.0.43-7.0.109; no public exploit code has been identified at time of analysis and Apache has released fixed versions 11.0.26, 10.1.60 and 9.0.122.
Cross-application authentication realm leakage in Apache Tomcat's Jakarta Authentication integration causes requests to be validated against the realm established by whichever co-hosted web application authenticated first, breaking per-application authentication boundaries. The flaw (CWE-287, improper authentication) affects Tomcat 11.0.0-M1 through 11.0.25, 10.1.0-M1 through 10.1.59, and 9.0.0.M4 through 9.0.121, plus end-of-life 8.5.0 through 8.5.100. Exploitation is gated by a non-default deployment - SimpleAuthConfigProvider as the default Jakarta Authentication provider with two or more applications sharing it - and the independent assessment rates the vector CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:L/A:L (unauthenticated, high attack complexity, low impact); no public exploit identified at time of analysis, the issue is not in CISA KEV, and the vulnerable configuration is not Tomcat's default.
Code execution in GNU Emacs versions prior to 31.2 occurs when Flymake performs on-the-fly syntax checking with a language backend other than Emacs Lisp, allowing code embedded in an untrusted file to run with the privileges of the user who opened or edited it. The CVSS vector (AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H) reflects that the attack is local and requires the victim to open a malicious file, but needs no privileges and can fully compromise confidentiality, integrity, and availability of that user's session; the flaw is tracked by Red Hat as CVE-2026-96442 with a vendor score of 7.8. No public exploit code and no confirmed active exploitation were identified at time of analysis, and the risk is concentrated in developer and administration workstations where Emacs commonly opens untrusted source files from downloads, mail attachments, or repositories.
Symlink-based path traversal in the `tar` source plugin of Apache BuildStream (versions up to and including 2.8.0, running on Python < 3.12) lets a malicious source tarball write arbitrary files on the host with the privileges of the user running BuildStream, resulting in full confidentiality, integrity, and availability impact to that user's environment. The vendor-published CVSS 3.1 base score is 9.8 (AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H), but the assessed vector adds UI:R: exploitation requires the victim to fetch and extract an attacker-controlled or otherwise untrusted source tarball, and it is not exploitable against projects that use only trusted, tracked sources, since SHA256 hash pinning of tracked tarballs prevents MITM substitution. No public exploit code and no confirmed active exploitation were identified at time of analysis (CISA SSVC exploitation: none; EPSS 0.19%, 10th percentile); the vendor-released patch is BuildStream 2.8.1, and hosts running Python >= 3.12 with BuildStream >= 2.3.0 are already protected because the Python `tarfile` extraction filter blocks the symlink escape.
Denial of service in FasterXML jackson-databind allows unauthenticated remote attackers to exhaust CPU resources by submitting crafted JSON that triggers quadratic complexity in forward-reference resolution for @JsonIdentityInfo object IDs. The attack requires the application to deserialize attacker-influenced JSON into a collection or map configured with @JsonIdentityInfo; a shallow document with N forward references defined in reverse order causes O(N^2) identity comparisons during deserialization. This is a conditional availability risk-only applications using identity-enabled deserialization on untrusted input are exposed-and no public exploit code has been identified at time of analysis, though a vendor patch is available.
Memory exhaustion in FasterXML jackson-databind allows unauthenticated remote attackers to grow an internal deserializer cache without bound by submitting JSON payloads that carry ever-changing, unrecognized polymorphic type IDs. TypeDeserializerBase caches each fallback resolution keyed by the raw attacker-supplied ID string, the map has no configurable size limit, and it lives for the lifetime of the long-lived ObjectMapper, so distinct unknown IDs translate directly into permanently retained heap - the reporter measured 10,000 retained entries from 10,000 distinct unknown IDs versus a single entry for 10,000 repetitions of one unknown ID. Exploitation requires a non-default configuration (name-based polymorphism via @JsonTypeInfo(use = Id.NAME) with a defaultImpl/fallback deserializer, an endpoint that accepts attacker-influenced type IDs, and a shared application-scoped mapper), so the affected population is substantially narrower than the 7.5 base score suggests; no public exploit has been identified at time of analysis and a vendor patch is available.
Denial of service in FasterXML jackson-core 2.8.0 and later: an unauthenticated remote attacker who can supply a malformed JSON token to a parser created via JsonFactory.createParser(DataInput) causes UTF8DataInputJsonParser._reportInvalidToken() to accumulate the entire offending token into an exception message without ever applying ErrorReportConfiguration.getMaxErrorTokenLength(). The reporter measured a 20,000,109-character exception message from a 20-million-character malformed token (versus 367 characters for identical input on the InputStream path), and because the StringBuilder also incurs byte-to-char expansion and internal array doubling, scaling the payload can exhaust heap and trigger a JVM-wide OutOfMemoryError. Only the DataInput parser path is reachable, impact is availability-only, and no public exploit identified at time of analysis; the 7.5 base score captures the theoretical worst case rather than the narrow real-world reachability.
Use-after-free in Google Chrome desktop versions prior to the September 2026 stable channel update allows remote attackers to cause information disclosure and memory corruption when a victim loads attacker-controlled web content. Exploitation requires user interaction (UI:R) and a high-complexity timing or heap-state precondition (AC:H), but no authentication (PR:N); the assessed CVSS vector is AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:H, yielding a base score of 8.3. A vendor patch is available in the September 2026 stable channel update, and no public exploit code or active exploitation has been identified in the provided intelligence.
A use-after-free in Google Chrome desktop can be triggered when a victim renders attacker-controlled web content, allowing memory corruption that may lead to code execution or disclosure of sensitive in-process data. Affected are Chrome desktop builds on Windows, macOS, and Linux prior to the September 2026 stable channel update; the CVSS vector (AV:N/AC:L/PR:N/UI:R/S:C/C:H/I:H/A:H, 9.6) requires no authentication or privileges but does require the victim to open or load a malicious or compromised page, and the S:C metric indicates the attack can affect a component beyond the one that is corrupted. Vendor-released patch available; no public exploit identified at time of analysis, and the reporting originator is Ubuntu.
Ubuntu Linux is the only product attributable to CVE-2026-95319, and the sole intelligence signal available is a vendor:ubuntu report with no accompanying advisory text, version range, or affected component. No description, CVSS score, CVSS vector, CWE classification, CPE string, EPSS score, KEV status, or reference link was supplied, so the vulnerability class, attack vector, privilege requirements, and impact cannot be characterized. No public exploit code has been identified at time of analysis, but this reflects absent data rather than a confirmed absence of exploitation, and the entry must be treated as unassessable until the upstream NVD or Ubuntu security advisory is retrieved.
CVE-2026-95318 is currently an unenriched record that cannot be meaningfully risk-assessed: the only attributable data point is that it was reported by vendor:ubuntu, with no description, no CPE strings, no CWE, no CVSS score or vector, and no references supplied. There is no confirmation of what component, package, or Ubuntu release is affected, whether exploitation requires authentication or user interaction, or what the resulting impact would be. Neither confirmed active exploitation (CISA KEV) nor publicly available exploit code exists in the available data, so the exploitation status is best described as no public exploit identified at time of analysis. This analysis should be treated as provisional and re-run once Ubuntu Security Notices or NVD enrichment supply a description, affected version range, and CVSS vector.
Use-after-free memory corruption in Google Chrome before 154.0.8037.57 allows a remote, unauthenticated attacker to potentially escape the browser's renderer sandbox by persuading a victim to visit attacker-controlled web content. Exploitation requires mandatory user interaction (UI:R) and a high-complexity precondition such as precise heap grooming or a timing race (AC:H), and no public exploit has been identified at time of analysis. Google has issued a vendor-released patch in version 154.0.8037.57 that removes the vulnerable code path.
Out-of-bounds write (CWE-787) in Google Chrome permits a remote, unauthenticated attacker to corrupt memory in the browser when a victim opens or is redirected to attacker-controlled web content (UI:R), with high confidentiality and integrity impact and a scope change (S:C) that the independent assessment reads as the corruption crossing the renderer sandbox boundary. Exploitation is not turnkey: the high attack complexity (AC:H) reflects a precondition the attacker must satisfy - typically heap grooming, a race, or a specific object/memory state - before the write is reachable, which constrains reliability and mass exploitation. No public exploit identified at time of analysis, and the fixed build 154.0.8037.57 is already shipping, so exposure is limited to un-updated Chrome clients, making this a genuine high-priority browser issue rather than a top-tier emergency.
Risk cannot be meaningfully assessed from the supplied data: CVE-2026-95283 is recorded only as a vendor-reported entry from Ubuntu, with the description marked '[Unknown description]' and no CVSS score, CVSS vector, CWE, CPE, or affected source package provided. Because the impacted component, deployment mode, and attack surface are all unidentified, no statement can be made about what an attacker could achieve, which versions are affected, or what preconditions (authentication, user interaction, configuration) would be required. There is no KEV listing and no public exploit code identified at time of analysis, so the entry should be treated as an unassessed tracking record rather than a confirmed exploitable issue.
Race condition in Google Chrome's stable channel allows a remote, unauthenticated attacker who persuades a user to interact with crafted content to win a time-of-check/time-of-use window and affect data beyond the vulnerable component's security scope. The CVSS 3.1 base score of 8.0 (AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:N) reflects high confidentiality and integrity impact with no availability impact, tempered by high attack complexity; no public exploit code or CISA KEV entry was identified at time of analysis. The flaw is addressed in the 154.0.8037.57 stable channel update, and while the vendor feed tags it Information Disclosure, the scoring indicates a broader boundary-crossing defect, so this is best treated as a moderately high-priority browser issue with partly inferred impact.
No technical details are available for CVE-2026-95338: the description field is empty and no CVSS score, CVSS vector, CWE classification, CPE string, or affected package name was supplied alongside the record. The only identifying signal is that the issue was reported under the vendor:ubuntu source, indicating it is tracked as a vulnerability in a package distributed through Ubuntu's repositories and handled via the Ubuntu Security Notice process. Risk cannot be meaningfully assessed from the available data, and there is no public exploit identified at time of analysis; exploitation prerequisites, impact, and affected versions must be obtained from the Ubuntu Security Notice for this CVE before any triage decision is made.
Improper input validation in Google Chrome before 154.0.8037.57 allows unauthenticated remote attackers to achieve high confidentiality, integrity, and availability impact with a scope change, though exploitation requires the victim to visit attacker-controlled web content and overcome high attack complexity. The flaw, tracked as CVE-2026-95341 with a CVSS 8.3 base score, affects default installations of Chrome on all platforms and could permit renderer-level code execution or sandbox escape; no public exploit identified at time of analysis. Vendor-released patch 154.0.8037.57 addresses the issue.
Type confusion in Google Chrome builds prior to 154.0.8037.57 lets a remote attacker who convinces a victim to load attacker-controlled web content corrupt memory inside the renderer and potentially achieve code execution in that process. The flaw is rated CVSS 8.8 (AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H, CWE-843), meaning no authentication or privileges are required but exploitation is not zero-click: the victim must render a malicious page, and because scope is unchanged the initial compromise is confined to the renderer sandbox unless chained with a separate sandbox escape. No public exploit code has been identified at time of analysis and the issue is not listed as confirmed actively exploited (CISA KEV); a vendor-released patch is available in Chrome 154.0.8037.57.
CVE-2026-95277 is currently an effectively empty record: no description, CVSS score, CVSS vector, CWE mapping, affected product, version range, or CPE data has been published, and the only intelligence signal is attribution to the Ubuntu vendor reporting channel. Because the vulnerability class, attack vector, authentication requirements, and impacted component are all unknown, risk cannot be meaningfully assessed from the available data and no exploit status can be confirmed - there is no public exploit identified at time of analysis, and the entry is not listed as confirmed actively exploited (CISA KEV). Organizations should treat this identifier as a placeholder to be re-triaged once Ubuntu publishes a corresponding security notice (USN) or the upstream CVE record is populated.
An out-of-bounds memory write in Google Chrome allows an unauthenticated, remote attacker to corrupt memory and potentially execute arbitrary code or crash the browser when a victim opens a crafted web page, as reflected by the CVSS 3.1 vector AV:N/AC:L/PR:N/UI:R/S:C/C:H/I:H/A:H (9.6, scope-changed). The flaw is classified as CWE-787 (out-of-bounds write) and is addressed by the vendor's stable channel update for desktop, with the fix credits originating from the Ubuntu vendor reporting channel, meaning downstream Chromium packages are also in scope. At time of analysis there is no public exploit identified, no CISA KEV listing, and no EPSS data provided; however, because browser memory-corruption bugs of this class are historically attractive to exploit developers, this remains a high-priority patching item despite the absence of exploitation telemetry.
Improper input validation in Google Chrome before 154.0.8037.57 can be triggered when a victim loads attacker-controlled web content, yielding a scope-changed impact rated high for confidentiality, integrity, and availability (CVSS 8.3). Exploitation is network-reachable without credentials (PR:N) but demands victim interaction (UI:R) plus non-trivial preconditions such as specific memory state, heap grooming, or timing windows (AC:H), and the affected component is undisclosed because chromium issue 517730821 is access-restricted. No public exploit code and no confirmed active exploitation were identified at time of analysis; a vendor patch is available in the Chrome stable channel at 154.0.8037.57.
Use-after-free in Google Chrome before 154.0.8037.57 can be triggered when a victim visits attacker-controlled web content, potentially allowing a remote unauthenticated attacker to corrupt memory and escape the renderer sandbox to execute code or disclose information. The vendor-assigned CVSS 3.1 score is 8.3 (AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:H), reflecting mandatory user interaction and high attack complexity. Vendor-released patch: 154.0.8037.57; no public exploit was identified at time of analysis.
Use-after-free memory corruption in Google Chrome desktop builds prior to 154.0.8037.57 allows remote attackers to achieve high-integrity and high-confidentiality impact against a victim who loads attacker-controlled web content and performs some interaction with it, but exploitation is not trivially repeatable: with AC:H it depends on winning a race or reliably grooming heap layout rather than a single deterministic trigger. The flaw is reachable without authentication (PR:N) and the vector's scope change (S:C) indicates the assessed impact crosses a component boundary, yet no public exploit code has been identified at time of analysis and there is no CISA KEV listing, so this should be treated as a high-priority patching item rather than a confirmed in-the-wild threat. A vendor patch is available.