Veille CVE

Semaine du 12 octobre au 18 octobre 2026

416 CVE

Critique 23 Élevée 165 Moyenne 184 Faible 19 Non évaluée 25 Exploitées (KEV) Tri : plus récentes « cli » ✕
  1. CVE-2026-98844 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: netfs: clear post-EOF pagecache when extending a file via write Fix netfs to erase the contents of a hole created after the EOF by an ordinary write if dirty data has been previously left there by writes through an mmapped region. Neither the buffered nor the unbuffered/DIO write path clears that stale pagecache. Zero the tail of the folio straddling the EOF before an extending write. That is the only folio that can hold data written past the EOF through an mmap, as pages wholly beyond the EOF can't be faulted in. The folio is zeroed rather than dropped so a concurrent extending write can't lose data. Both write paths downgrade the i_rwsem to shared, so extending writes can run concurrently and the i_size read by the caller may be stale by the time the folio is locked. Re-read i_size under the folio lock and clamp the zeroed range up to it, so a racing write that already put data into the folio isn't clobbered. Wait for any writeback on the folio to finish before zeroing it so that the pagecache isn't modified while it may still be read by the transport during transmission. Honour IOCB_NOWAIT by returning -EAGAIN rather than blocking on the folio lock, on writeback, or in folio_mkclean()'s rmap walk when the folio is mapped. truncate_pagecache() can't be used here: it must be called with the i_rwsem held exclusively, but these write paths only hold it shared, and it would block unconditionally, breaking IOCB_NOWAIT. Callers that hold i_rwsem exclusively for the whole resize (truncate, setattr, fallocate, clone) exclude any genuine concurrent buffered writer, so staleness can instead be decided from the folio's dirty state, as pagecache_isize_extended() already does for filesystems that serialise writes against truncate/setattr via a single i_rwsem. Export netfs_clear_stale_post_isize() helper to handle such case. The helper is required by the CIFS client to fix generic/363.

  2. CVE-2026-98839 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: vt: selection: Fix unsigned underflow and slab-out-of-bounds read in paste_selection() In paste_selection(), the loop copies min_t(unsigned int, vc_sel.buf_len - pasted,tty->receive_room) bytes per iteration into tty_ldisc_receive_buf() and increments pasted += count. Because the selection mutex (vc_sel.lock) is dropped inside the loop Whenever the line discipline buffer fills up and paste_selection() sleeps on tty->write_wait, a concurrent TIOCLINUX (TIOCL_SETSEL) ioctl can replace vc_sel.buffer with a shorter selection and reduce vc_sel.buf_len below pasted. When paste_selection() resumes, vc_sel.buf_len - pasted underflows as an unsigned integer to a large positive value, causing tty_ldisc_receive_buf(ld, vc_sel.buffer + pasted, NULL, count) to read up to 4094 bytes out-of-bounds past the newly allocated vc_sel.buffer. Fix this by terminating the loop when pasted >= vc_sel.buf_len. Kernel stack trace (Linux 7.3.0-rc3): ================================================================== BUG: KASAN: slab-out-of-bounds in n_tty_receive_buf_common+0xa01/0x1650 Read of size 4094 at addr ffff888101c58010 by task kworker/u17:1/65 Workqueue: events_unbound flush_to_ldisc Call Trace: <TASK> dump_stack_lvl+0x70/0xa0 print_report+0x153/0x4c6 kasan_report+0xf1/0x120 kasan_check_range+0x11c/0x200 __asan_memcpy+0x29/0x70 n_tty_receive_buf_common+0xa01/0x1650 tty_ldisc_receive_buf+0x66/0x110 tty_port_default_receive_buf+0x6b/0xb0 flush_to_ldisc+0x1b4/0x410 process_one_work+0x6ff/0x1110 worker_thread+0x4a8/0xb70 kthread+0x307/0x3e0 ret_from_fork+0x3ed/0x680 </TASK> ==================================================================

  3. CVE-2026-98758 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: usb: typec: ucsi: glink: fix use-after-free of ucsi on remove The pmic_glink client is released by the devres cleanup, which runs after pmic_glink_ucsi_remove() has returned, so its notification callbacks can queue work until then. Work that runs after ucsi_unregister() has been called touches freed state: it dereferences the connector array that ucsi_unregister() freed, or registers and unregisters the instance a second time. Disable both work items with disable_work_sync() before unregistering, and unregister only if the instance is still registered. This issue was found by an in-house static analysis tool.

  4. CVE-2026-98742 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: ALSA: seq: Serialize compat port-info ioctls The native sequencer ioctl path serializes handler calls with client->ioctl_mutex, but the translated port-info compat path invokes the same handlers through snd_seq_kernel_client_ctl() without taking that mutex. This lets concurrent compat CREATE_PORT requests pass the port-count check before any request reaches the serialized insertion. The computed integer port index can then exceed the address field range and wrap to an existing index. Subsequent subscriber teardown can resolve the duplicate address to the wrong port and access a freed subscriber. Take ioctl_mutex while dispatching converted port-info requests, matching the native ioctl path. All translated port-info commands share this helper, so their accesses to the client port state are serialized as well.

  5. CVE-2026-98686 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: smb: client: discard post-EOF pagecache when extending a file via zero range smb3_zero_range() removed the pagecache only for the range being zeroed, so when it extends the file the folio straddling the old EOF - which may hold data dirtied past EOF through an mmap - was never dropped and became visible as file content once the file was extended. Lower the discard boundary to the smaller of the zero-range offset and the old EOF.

  6. CVE-2026-98685 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: smb: client: discard post-EOF pagecache when extending a file via clone range cifs_remap_file_range() discarded the pagecache only for the folios overlapping the destination region, so when a clone extends the file the folio straddling the old EOF - which may hold data dirtied past EOF through an mmap - was never dropped and became visible as file content once the file was extended. Start the discard at the smaller of fstart and the old EOF. i_size is still old here, so nothing past the old EOF is flushed to the server.

  7. CVE-2026-98684 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: smb: client: flush and commit data before querying allocated ranges The mode 0 / FALLOC_FL_KEEP_SIZE fallocate emulation for small interior ranges in smb3_simple_fallocate_range() issues FSCTL_QUERY_ALLOCATED_RANGES and zero-fills any sub-range the server reports as an unallocated hole, to force block allocation without changing file contents. This is unsafe against Windows servers. Per the Win32 documentation, FSCTL_QUERY_ALLOCATED_RANGES only reports ranges that *may* contain nonzero data and is explicitly not coherent with recently written data: "a call to FSCTL_QUERY_ALLOCATED_RANGES [after writing to a network file] would not necessarily return a correct list of allocated regions. To ensure coherency [...] flush the data to the file" fsx (generic/363) reproduces the resulting corruption against Windows Server 2022: a range is written, an uncached read returns the data, yet the next FSCTL_QUERY_ALLOCATED_RANGES on that range reports it as a whole hole, so the emulation overwrites live data with zeroes. A network trace confirmed the WRITE, the READ returning data and the query returning an empty range list within microseconds. Samba does not exhibit this; its query is coherent with the written data. Write back the dirty pages, drain any outstanding I/O and issue an SMB2 FLUSH to force the server to commit the data before querying the allocated ranges, so the query reflects the data actually present. With this, generic/363 passes against both Windows Server 2022 and Samba over 100000 fsx operations.

  8. CVE-2026-98683 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: smb: client: flush dirty data before zeroing a range smb3_zero_range() emulates FALLOC_FL_ZERO_RANGE by invalidating the page cache over the target range with truncate_pagecache_range() and then issuing FSCTL_SET_ZERO_DATA to the server. Dirty data was only flushed conditionally, when the range reached or extended EOF. For a purely interior zero range that does not reach EOF, no flush happened, so a dirty folio overlapping the range could be written back to the server after the FSCTL and refill the range that was just zeroed with stale data. Fix this by unconditionally flushing and waiting for dirty data in the range before invalidating the page cache and issuing FSCTL_SET_ZERO_DATA, exactly as was done for smb3_punch_hole() in commit d7d2adcd022b ("smb/client: flush dirty data before punching a hole"); the two paths are structurally identical here. This is observed as generic/363 randomly reading stale data where a zeroed range is expected against Windows Server.

  9. CVE-2026-98682 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: smb: client: drain and invalidate before server-side copy/clone cifs_file_copychunk_range() and the clone (FICLONE) path of cifs_remap_file_range() do not serialise the destination page cache against the server-side copy the way the other server-side range operations (smb3_zero_range(), smb3_punch_hole(), smb3_collapse_range() and smb3_insert_range()) do. cifs_file_copychunk_range() invalidates the destination range with filemap_invalidate_inode(), which takes and drops the mapping's invalidate_lock internally, so the lock is no longer held when the copychunk ioctl is issued. It also never drains in-flight netfs I/O on the target. The clone path never takes the invalidate_lock at all (only i_rwsem), uses a bare truncate_inode_pages_range() and likewise does not drain outstanding I/O. As a result an asynchronous destination writeback can complete after the server-side copy/clone has run and reinstate stale data over the region just written by the server, corrupting the file. This is the same class of corruption as commit d7d2adcd022b ("smb/client: flush dirty data before punching a hole") and has been seen randomly in generic/363 against Windows Server. Fix both paths to follow the established ordering: hold the target mapping's invalidate_lock across the flush and invalidation of the destination and the server ioctl, and call netfs_wait_for_outstanding_io() on the target to drain in-flight writes before the ioctl is issued. Only the target inode's invalidate_lock is required, as the source is merely flushed and not invalidated; i_rwsem (already held via lock_two_nondirectories()) is acquired before the invalidate_lock, matching the VFS lock ordering. Since filemap_invalidate_inode() takes that same lock internally, replace it with its own unmap/flush/invalidate steps instead of calling it, and return early on a zero-length copy to avoid a range underflow.

  10. CVE-2026-98681 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: smb: client: split cifsFileInfo bitfields to avoid shared-byte RMW races The invalidHandle, swapfile, oplock_break_cancelled, offload, and status_file_deleted fields are stored in the same bitfield byte in struct cifsFileInfo, but are updated in different code paths that may run simultaneously, and are protected by different locks. Since bitfield assignments generate byte-level read-modify-write operations, a modification to one flag can overwrite a concurrent modification to another flag. To avoid these races, convert these flags from a bitfield to separate bool fields.

  11. CVE-2026-98613 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: bpf: Bound ownership depth through local kptrs and graph roots Program-allocated objects can own other local objects through referenced kptrs. bpf_obj_free_fields() follows those pointers through __bpf_obj_drop_impl() synchronously, before the object storage is freed through RCU. A self-referential local kptr type therefore permits arbitrarily deep object chains, and dropping the head can exhaust the kernel stack. Long acyclic type chains have the same problem. btf_check_and_fixup_fields() still assumes referenced kptrs only point to kernel types and checks ownership through list and rbtree roots only. Its existing rule is sufficient for graph-only cycles: the target of each graph edge must contain a node, so every type in a cycle has both a root and a node. The rule rejects such a type owning another root, breaking every cycle. It also limits graph-only chains to three types, or two if the first type contains a node, and conservatively rejects longer acyclic chains. The missing local-kptr edges, rather than a missed graph-only cycle, are the bug introduced by support for bpf_kptr_xchg() into local kptrs. Replace that restriction with one bounded ownership walk covering graph roots and local referenced kptrs. Run it after all BTF records have been fixed up, reject cycles and paths deeper than eight record-bearing types, and cache each type's suffix depth while checking it against the remaining budget. This also permits the longer acyclic graph-only layouts rejected by the old rule; update their existing BTF tests accordingly. Keep the bound independent of MAX_CALL_FRAMES because recursive destruction can run below a BPF call chain. A plain local pointee without special-field metadata adds only a final non-recursing drop. Non-owning kptrs and kernel-BTF kptrs do not recurse through local records and remain outside the walk. Include local percpu-kptr edges too, although allocation of percpu objects with special fields is currently forbidden, so that relaxing that restriction cannot bypass the ownership bound. btf_check_and_fixup_fields() continues to initialize graph_root.value_rec, including for separately allocated map records. The ownership relationships belong to immutable program BTF and only need validation at BTF load time.

  12. CVE-2026-98587 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: firewire: cdev: fix back-transition for iso_resource_auto client resource The todo member of iso_resource_auto structure represents the state of the client resource and normally transitions in the following order: ISO_RES_AUTO_ALLOC -> ISO_RES_AUTO_REALLOC -> ISO_RES_AUTO_DEALLOC However, concurrent access from the work item and the file descriptor release function can cause the state to transition backwards from ISO_RES_AUTO_DEALLOC to ISO_RES_AUTO_REALLOC. Prevent the back-transition by checking the current state before updating it in the work item.

  13. CVE-2026-98571 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: net: stmmac: clear stale buf->page after recycling on skb build failure In stmmac_rx(), when napi_build_skb() fails the descriptor page is recycled back to the page pool with page_pool_recycle_direct(), but buf->page is left pointing at the recycled page, unlike every other consumption site in the function which clears the pointer after handing the page away. With the stale pointer stmmac_rx_refill() skips the replacement allocation and programs the already-recycled page back into the RX descriptor. Clear buf->page on the napi_build_skb() failure path to keep the buffer lifecycle consistent with the other consumption sites.

  14. CVE-2026-98523 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: drm/client: fix restore of partially initialized client I got a null-ptr-deref report when closing a DRM file descriptor: WARNING: drivers/gpu/drm/drm_atomic.c:2031 at __drm_atomic_helper_set_config+0x18e/0x1b0 [drm] Call Trace: drm_client_modeset_commit_atomic+0x16b/0x220 [drm] drm_client_modeset_commit_locked+0x56/0x160 [drm] drm_client_modeset_commit+0x21/0x40 [drm] __drm_fb_helper_restore_fbdev_mode_unlocked.part.0+0x7b/0x80 drm_fbdev_client_restore+0xe/0x20 [drm_client_lib] drm_client_dev_restore+0x9f/0xc0 [drm] drm_release+0xc5/0xe0 [drm] The warning is followed by a NULL pointer dereference: BUG: kernel NULL pointer dereference, address: 0000000000000008 RIP: __drm_fb_helper_restore_fbdev_mode_unlocked.part.0+0x41/0x80 [drm_kms_helper] Call Trace: drm_fbdev_client_restore+0xe/0x20 [drm_client_lib] drm_client_dev_restore+0x9f/0xc0 [drm] drm_release+0xc5/0xe0 [drm] __fput+0xdc/0x2b0 __x64_sys_close+0x39/0x80 do_syscall_64+0x8d/0x460 entry_SYSCALL_64_after_hwframe+0x76/0x7e drm_client_register() adds the DRM client to the device client list before invoking the initial hotplug callback. If the hotplug callback fails, the client remains registered. For the fbdev client, a failure during drm_fb_helper_initial_config() causes the partially initialized fbdev helper to be cleaned up. drm_fb_helper_fini() releases fb_helper->info and leaves it NULL. The fbdev client therefore remains registered even though there is no fully initialized framebuffer device. Later, when userspace closes the DRM file descriptor, drm_release() can invoke the restore callbacks of registered DRM clients: drm_release() drm_client_dev_restore() drm_fbdev_client_restore() drm_fb_helper_restore_fbdev_mode_unlocked() drm_fbdev_client_restore() currently restores the fbdev state unconditionally. For a partially initialized fbdev client this can submit an incomplete modeset state and subsequently access fbdev state which has not been initialized, resulting in the warning and NULL pointer dereference above. drm_fbdev_client_unregister() already uses fb_helper->info to distinguish a fully probed framebuffer device from a partially initialized client. Use the same condition in drm_fbdev_client_restore() and skip restore if no framebuffer device has been successfully initialized.

  15. CVE-2026-98515 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: drm/nouveau: fix double-free in nvif_vmm_dtor On failure, nouveau_cli_init() calls nouveau_cli_fini() to tear the client down. Then, nouveau_drm_open() also enters into its cleanup path and calls nouveau_cli_fini() AGAIN. nouveau_cli_fini() calls nouveau_vmm_fini(): void nouveau_vmm_fini(struct nouveau_vmm *vmm) { nouveau_svmm_fini(&vmm->svmm); nvif_vmm_dtor(&vmm->vmm); vmm->cli = NULL; } Inside nvif_vmm_dtor(), vmm->page is freed unconditionally: void nvif_vmm_dtor(struct nvif_vmm *vmm) { kfree(vmm->page); nvif_object_dtor(&vmm->object); } vmm->page is never cleared after being freed, so the second call of nvif_vmm_dtor() will cause a double-free. Found by fuzzing the nouveau driver with a modified Syzkaller: BUG: KASAN: double-free in nvif_vmm_dtor+0x31/0x50 drivers/gpu/drm/nouveau/nvif/vmm.c:194 Free of addr ffff888010fcdc30 by task syz.0.173/2567 CPU: 1 UID: 0 PID: 2567 Comm: syz.0.173 Not tainted 7.2.0 #24 PREEMPT(lazy) Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-debian-1.16.3-2 04/01/2014 Call Trace: <TASK> __dump_stack lib/dump_stack.c:94 [inline] dump_stack_lvl+0x95/0xe0 lib/dump_stack.c:120 print_address_description mm/kasan/report.c:378 [inline] print_report+0xcb/0x5a0 mm/kasan/report.c:482 kasan_report_invalid_free+0xaa/0xd0 mm/kasan/report.c:557 check_slab_allocation+0xe4/0x110 mm/kasan/common.c:235 kasan_slab_pre_free include/linux/kasan.h:199 [inline] slab_free_hook mm/slub.c:2622 [inline] slab_free mm/slub.c:6377 [inline] kfree+0x192/0x590 mm/slub.c:6692 nvif_vmm_dtor+0x31/0x50 drivers/gpu/drm/nouveau/nvif/vmm.c:194 nouveau_vmm_fini+0x16/0x50 drivers/gpu/drm/nouveau/nouveau_vmm.c:127 nouveau_cli_fini+0x10e/0x210 drivers/gpu/drm/nouveau/nouveau_drm.c:225 nouveau_drm_open+0x24e/0x740 drivers/gpu/drm/nouveau/nouveau_drm.c:1255 drm_file_alloc+0x5f2/0xad0 drivers/gpu/drm/drm_file.c:176 drm_open_helper+0x1d7/0x4a0 drivers/gpu/drm/drm_file.c:335 drm_open+0x190/0x3d0 drivers/gpu/drm/drm_file.c:388 drm_stub_open+0x1f2/0x460 drivers/gpu/drm/drm_drv.c:1211 chrdev_open+0x21c/0x660 fs/char_dev.c:411 do_dentry_open+0x59d/0x12b0 fs/open.c:947 vfs_open+0x82/0x390 fs/open.c:1052 do_open fs/namei.c:4700 [inline] path_openat+0x2345/0x3420 fs/namei.c:4863 do_file_open+0x207/0x460 fs/namei.c:4892 do_sys_openat2+0xd1/0x1d0 fs/open.c:1368 do_sys_open fs/open.c:1374 [inline] __do_sys_openat fs/open.c:1390 [inline] __se_sys_openat fs/open.c:1385 [inline] __x64_sys_openat+0x144/0x200 fs/open.c:1385 do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline] do_syscall_64+0x115/0x690 arch/x86/entry/syscall_64.c:94 entry_SYSCALL_64_after_hwframe+0x77/0x7f RIP: 0033:0x7fc6d687594d Code: ff c3 66 2e 0f 1f 84 00 00 00 00 00 90 f3 0f 1e fa 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 b0 ff ff ff f7 d8 64 89 01 48 RSP: 002b:00007fc6d5295008 EFLAGS: 00000246 ORIG_RAX: 0000000000000101 RAX: ffffffffffffffda RBX: 00007fc6d6b06180 RCX: 00007fc6d687594d RDX: 0000000000022501 RSI: 0000200000000000 RDI: ffffffffffffff9c RBP: 00007fc6d691c303 R08: 0000000000000000 R09: 0000000000000000 R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000 R13: 00007fc6d6b06218 R14: 00007fc6d6b06180 R15: 00007ffd9451d760 </TASK> Allocated by task 2567 on cpu 1 at 163.593900s: kasan_save_stack+0x24/0x50 mm/kasan/common.c:57 kasan_save_track+0x17/0x60 mm/kasan/common.c:78 poison_kmalloc_redzone mm/kasan/common.c:398 [inline] __kasan_kmalloc+0xaa/0xb0 mm/kasan/common.c:415 kasan_kmalloc include/linux/kasan.h:263 [inline] __do_kmalloc_node mm/slub.c:5334 [inline] __kmalloc_noprof+0x304/0x7c0 mm/slub.c:5359 _kmalloc_noprof include/linux/slab.h:992 [inline] nvif_vmm_ctor+0x3c0/0x7e0 drivers/gpu/drm/nouveau/nvif/vmm.c:237 nouveau_vmm_init+0x40/0x90 drivers/gpu/drm/nouveau/nouveau_vmm.c:134 nouveau_cli_init+0x7b9/0 ---truncated---

  16. CVE-2026-98493 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: net: ipconfig: bound DHCP option construction ic_dhcp_init_options() appends the hostname (option 12), vendor-class (option 60) and client-ID (option 61) options into the fixed 312-byte bootp_pkt.exten[] buffer. Only the client-ID branch checked the remaining space; the hostname and vendor-class writes were unbounded. A 64-byte hostname together with the maximum 252-byte dhcpclass= identifier needs 18 + (2 + 64) + (2 + 252) = 338 of the 312 available bytes even before the terminating END marker, so the vendor-class memcpy runs past the end of exten[]. With CONFIG_FORTIFY_SOURCE this is reported as a field-spanning write and, when the kernel is booted with panic_on_warn=1, aborts boot with a panic. Route the optional options through a common helper that makes sure the option, its 2-byte header and the END marker all fit and drops an option that would not. Configurations with short options keep sending exactly the same bytes as before.

  17. CVE-2026-98454 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: smb: client: fix create context out-of-bounds reads smb2_parse_contexts() validates the complete create-context area but does not limit each record to its Next field before dispatching it. A malformed chain can therefore expose bytes beyond the current context to a handler. The QFid handler also used a full response-structure cast although it only reads DiskFileId. The SMB2/SMB3 lease parsers made the same layout assumption: they read LeaseState and LeaseFlags at canonical offsets rather than at DataOffset. A valid non-canonical DataOffset could therefore yield unrelated in-bounds data, while a short DataLength was still accepted. Limit each context to its Next value, reject offsets before the context header, and reject malformed chains. Bound the name range by the current context and do not dispatch a known handler when DataLength is zero. Read the QFid DiskFileId only when the context data covers that field. Parse the lease context from DataOffset and require DataLength to match the v1 or v2 lease_context size used by ksmbd. A size mismatch skips lease parsing without failing the open.

  18. CVE-2026-98453 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: smb: client: close handle after create-context parsing failure SMB2_open() accounts a successful CREATE response as a remote open before parsing its create contexts. If smb2_parse_contexts() rejects malformed context data, SMB2_open() returns without closing the handle, leaving the server-side handle open and num_remote_opens elevated. Close the handle after a post-CREATE context parsing failure so the error path releases the remote resource and balances the open count.

  19. CVE-2026-98452 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: smb: client: clean up failed cached directory opens open_cached_dir() sends CREATE and QUERY_INFO as a compound request. If the CREATE succeeds but a later command returns an error, the function must retain the CREATE FID so common cleanup can issue SMB2_close(). It also must not treat a response error as a valid CREATE. Validate the CREATE response before using its fields, record the FIDs, and mark the handle open before handling errors from later compound commands. Move the -EREMCHG reconnect handling before response validation so a missing response does not hide the reconnect request. Count the handle when it is marked open; confirmed close responses decrement the counter, while existing close retry behavior remains best effort on transport failures.

  20. CVE-2026-98451 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: smb: client: use finish_no_open() for non-regular inodes An O_CREAT open can find an existing symlink or another non-regular inode. cifs_atomic_open() calls finish_open() on it and attaches a cifsFileInfo. Symlink inodes have no CIFS release operation, so the dentry reference held by cifsFileInfo is leaked. FMODE_OPENED also prevents the VFS from following the symlink. Track whether cifs_do_create() returned an open server handle. For non-regular inodes, close the handle if present, remove the pending open, and call finish_no_open() so the VFS can continue the lookup. Do not set FMODE_CREATED unless a regular file was opened. For O_NOFOLLOW with __O_REGULAR, return -ELOOP before the VFS's -EFTYPE check. Defer closing a legacy POSIX handle on a non-regular inode until after inode lookup. This avoids closing it again if lookup fails.

  21. CVE-2026-98450 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: smb: client: close completed creates on compound wait errors compound_send_recv() waits for responses in order. If a later wait is interrupted, or if a later MID fails during response synchronization, an earlier CREATE may already have opened a remote handle. The earlier mid is then released without invoking handle_cancelled_mid(), leaving the remote handle open because no FID was copied to the caller. Mark completed earlier mids as cancelled when a compound wait or MID synchronization aborts. Keep their response buffers attached while the MIDs are synchronized, and transfer them only after synchronization of the processed responses, so the release path can inspect successful CREATE responses and queue SMB2_close() after a later failure. Account for a remote open only after the close work is allocated and before it is queued, since the caller has not yet updated num_remote_opens. Mark the create+close compound used by smb2_unlink() so it is not closed again. Non-CREATE responses and compounds that already include a close keep their existing behavior.

  22. CVE-2026-98449 – Non évaluée

    Linux

    In the Linux kernel, the following vulnerability has been resolved: smb: client: delete compound mids on send failure before unlock When sending a compound request fails, smb_send_rqst() kicks off a reconnect. compound_send_recv() still has those mids on pending_mid_q, but it unlocks the server without removing them first. During reconnect, cifs_abort_connection() walks pending_mid_q and runs each mid callback. With no response yet, those callbacks return credits and drop in_flight. Then compound_send_recv()'s send-error path returns the same credits again. in_flight ends up decremented twice and smb2_add_credits() WARNs. syzbot hits this during SMB2_negotiate when the socket send fails. cifs_call_async() already calls delete_mid() before unlock on send failure. Do the same for compound chains and set cancelled_mid[] so the out: path does not delete them again.

  23. CVE-2026-112063 5.9 Moyenne

    strophe libstrophe

    libstrophe through 0.14.0 contains a stack-based buffer overflow vulnerability in SCRAM_Hi that allows malicious XMPP servers to overflow a 128-byte stack buffer by sending an oversized SCRAM salt. Attackers can send a server-first-message whose s= attribute decodes to more than 124 bytes, aborting the client or overwriting stack memory in NDEBUG builds.

  24. CVE-2026-112061 4.3 Moyenne

    RTMPDump librtmp

    librtmp (RTMPDump) through 2.6 contains a heap out-of-bounds read vulnerability in AMFProp_Decode that reads an AMF0 strict-array element count before checking that four bytes remain. A malicious RTMP server or man in the middle can send an AMF0 body ending with a strict-array type byte and fewer than four trailing bytes to crash the client.

  25. CVE-2026-112050 3.1 Faible

    gitify-app gitify

    Gitify before 7.10.0 contains a login cross-site request forgery vulnerability that allows remote attackers to bind a victim's client to their account because performGitHubWebOAuth() omits OAuth state validation. Attackers can get a gitify://oauth?code= link with their own authorization code opened during a victim's pending login, causing Gitify to display attacker-controlled notifications as the victim's own.

  26. CVE-2026-112005 7.7 Élevée

    clidey whodb

    WhoDB 0.104.0 before 0.135.0 contains a path traversal vulnerability that allows remote attackers to read and write arbitrary host files by running DuckDB filesystem functions through the query editor. Attackers can log in to a DuckDB file without credentials and use read_text, read_blob, COPY ... TO or ATTACH to steal session.key and stored database credentials or overwrite files.

  27. CVE-2026-110981 7.1 Élevée

    Unknown WP Photo Album Plus

    The WP Photo Album Plus WordPress plugin before 9.2.11.002 does not properly escape a category navigation value before reflecting it into a JavaScript event handler on its gallery breadcrumb links, allowing unauthenticated attackers to inject arbitrary script that executes in the browser of anyone who follows a crafted gallery link and clicks the breadcrumb, including a logged-in administrator.

  28. CVE-2026-111874 3.7 Faible

    OpenCATS

    A flaw has been found in OpenCATS 0.10.0. Affected is the function processLogin of the file modules/careers/CareersUI.php of the component Career Portal. This manipulation of the argument applyToJobSubAction causes improper restriction of excessive authentication attempts. Remote exploitation of the attack is possible. A high degree of complexity is needed for the attack. The exploitability is told to be difficult. The exploit has been published and may be used. Patch name: 1cf4fa3ee9054314b023be79d404078645c22ade. It is suggested to install a patch to address this issue. The project maintainer explains: "Candidates will now need to authenticate each time they access the Career Portal and successfully complete the CAPTCHA as part of that authentication flow. The authenticated candidate identity is now maintained server-side rather than relying on the previous persistent client-side cats…cw cookie mechanism. Application-level rate limiting is deliberately not in scope for this change. In my view, request throttling and brute-force protection of that kind are better applied at the network, reverse-proxy or web-server layer, where they can be configured appropriately for each deployment without OpenCATS having to implement and maintain its own rate-limiting framework. I'll update the online documentation to that effect."

  29. CVE-2026-111934 3.7 Faible

    coder

    Coder before 2.34.7 contains an IP spoofing vulnerability in the ExtractRealIPAddress middleware that trusts the leftmost, client-supplied X-Forwarded-For value instead of the client IP appended by a reverse proxy. Attackers can send forged X-Forwarded-For values through appending proxies to bypass per-IP rate limits on login, one-time passcode, and password validation endpoints, enabling brute-force attacks, and to falsify IP addresses recorded in audit logs.

  30. CVE-2026-111903 5.4 Moyenne

    thorsten phpMyFAQ

    phpMyFAQ 4.2.0-alpha before 4.2.0-beta contains a stored cross-site scripting vulnerability in SvgSanitizer that allows users with FAQ_EDIT permission to bypass sanitization using XML DTD custom entities. Attackers can upload an SVG defining an entity that expands to javascript in an anchor href, executing script when viewers click it to hijack admin sessions.

  31. CVE-2026-111881 7.6 Élevée

    FreeRDP

    FreeRDP from 3.14.1 before 3.32.0 contains an integer underflow in the WinPR BMP writer, where a mismatched stride calculation passes SIZE_MAX - 3 to Stream_Zero for sub-byte images. A malicious RDP clipboard peer can return a crafted 1-bpp BMP processed by clipboard_synthesize_image_format_to_bmp for CF_DIB/CF_DIBV5, crashing the client or causing an out-of-bounds write.

  32. CVE-2024-58398 4.3 Moyenne

    coder

    Coder 2.3.1 before 2.14.4, 2.15.x before 2.15.3, and 2.16.0 before 2.16.1 contains an open redirect vulnerability in the login page because the redirect query parameter is improperly sanitized. Attackers can craft a Coder login URL that, when clicked by a logged-in user, redirects the victim to an attacker-controlled website for phishing.

  33. CVE-2026-97314 6.5 Moyenne

    Unknown ThirstyAffiliates

    The ThirstyAffiliates WordPress plugin before 3.12.1 does not check the status or type of the object requested by its link redirect tracking handler, allowing unauthenticated attackers to disclose the destination URLs of unpublished affiliate links and to record fabricated click statistics.

  34. CVE-2026-92978 4.3 Moyenne

    wpeverest User Registration & Membership – Free & Paid Memberships, Subscriptions, Content Restriction, User Profile, Custom User Registration & Login Builder

    The User Registration & Membership – Free & Paid Memberships, Subscriptions, Content Restriction, User Profile, Custom User Registration & Login Builder plugin for WordPress is vulnerable to Cross-Site Request Forgery in all versions up to, and including, 5.2.8 This is due to missing or incorrect nonce validation on the trigger_bulk_action function. This makes it possible for unauthenticated attackers to force administrators to approve pending user registrations or deny arbitrary user accounts, locking targeted users out of the site via a forged request granted they can trick a site administrator into performing an action such as clicking on a link. An attacker can bypass the earlier nonce check in trigger_query_actions() by supplying an empty new_role parameter, which causes that handler to return early and cede control to the unprotected trigger_bulk_action() handler; the exit() call at the end of that handler additionally prevents WordPress core's own bulk-users nonce validation from executing.

  35. CVE-2026-111601 6.1 Moyenne

    hippooo Hippoo Mobile App for WooCommerce

    The Hippoo Mobile App for WooCommerce plugin for WordPress is vulnerable to Reflected Cross-Site Scripting via the 'token_id' parameter in all versions up to, and including, 1.12.0 due to insufficient input sanitization and output escaping. This makes it possible for unauthenticated attackers to inject arbitrary web scripts in pages that execute if they can successfully trick a user into performing an action such as clicking on a link. The URL path regex constrains the path segment to word characters, but WP_REST_Request's query-string parameter precedence allows a ?token_id= value to override it and reach the handler verbatim, bypassing the regex constraint entirely.

  36. CVE-2026-111044 7.6 Élevée

    LizardByte Sunshine

    Sunshine is a self-hosted game stream host for Moonlight. From 2026.329.165800 until 2026.906.222525, Sunshine accepted any client certificate that chained to a stored paired-client certificate, but it did not require the presented certificate to be the exact enabled paired identity. Certificate text was also compared without canonicalizing equivalent PEM encodings. A previously paired client that retains its private key can present an alternate certificate derived from its paired certificate and bypass the Disable control. Equivalent PEM encodings could also cause inconsistent client identity handling. A client that was never paired cannot use this issue, and removing the paired client removes its trust relationship. This issue is fixed in version 2026.906.222525.

  37. CVE-2026-96620 9.2 Critique

    anthropics claude-agent-sdk-python

    Claude Agent SDK is a Python SDK for Claude Code. Prior to 0.2.121, the resume option is passed to the Claude CLI as a separate token without validation, so a dash-leading value is parsed as an independent CLI flag. An attacker can inject the --mcp-config flag to define a stdio MCP server, causing arbitrary command execution at CLI startup outside the SDK permission controls. The issue is exploitable when an application passes untrusted input as the session ID to resume. This issue is fixed in version 0.2.121.

  38. CVE-2026-85977 5.3 Moyenne

    Red Hat Hardened Images · Red Hat Enterprise Linux 10 · Red Hat Enterprise Linux 6 · Red Hat Enterprise Linux 7

    A flaw was found in libtirpc. A remote attacker can exploit this vulnerability by sending specially crafted PMAPPROC_DUMP replies to a client application that uses the pmap_getmaps() function against an attacker-controlled host. This unbounded decoding can lead to excessive memory and CPU allocation, causing the client process to experience a Denial of Service (DoS) due to resource exhaustion.

  39. CVE-2026-111714 3.1 Faible

    AmauriC tarteaucitron.js

    tarteaucitron.js before 1.37.0 contains a regular expression denial of service vulnerability that allows attackers to block visitor pages by supplying unescaped cookie names to cookie.purge. Attackers can plant a cookie from a sibling subdomain or inject a purgeBtn element with a catastrophic-backtracking data-cookie pattern that freezes the page when the purge button is clicked.

  40. CVE-2026-111705 4.3 Moyenne

    fushengqian fuint

    Fuint through commit 57ec018 contains an insecure direct object reference vulnerability that allows authenticated members to read other members' booking records by supplying arbitrary bookId values to /clientApi/book/myBookDetail. Attackers can iterate the auto-increment bookId to retrieve contact names, mobile numbers, in-store verification codes, service dates, stores and remarks.

  41. CVE-2026-111697 6.5 Moyenne

    roinli HUIZHI-ChargeOS-cloud

    HUIZHI ChargeOS cloud through 3.0.9 contains an insecure direct object reference vulnerability in the hcp-mp MeController and OrderController that trusts client-supplied userId, openId and orderNumber parameters. Attackers can enumerate sequential member ids against endpoints like /me/getMemberBalanceByUserId and /order/orderDetail to read any member's wallet, credit, profile and charging orders.

  42. CVE-2026-111670 4.3 Moyenne

    clickvisual

    ClickVisual through 1.1.2 contains a missing authorization vulnerability in the /api/v2/reports handlers that allows authenticated low-privileged users to access ClickHouse instances they were not granted. Attackers can iterate the instance-id parameter in ReportSourceDatabases, ReportSourceTables, ReportTableColumns and WhereCheckRun to enumerate schemas and infer row counts of restricted tables.

  43. CVE-2026-111639 8.1 Élevée

    mlogclub bbs-go

    bbs-go through 4.4.6 contains an authentication bypass vulnerability in the /api/login/google_one_tap endpoint that allows remote attackers to log in as other users by submitting Google ID tokens issued to any OAuth client. Attackers operating their own Google OAuth client can capture a victim's Google-signed ID token and replay it, since the aud claim is never checked, to obtain an authenticated session as the victim's Google-linked account.

  44. CVE-2026-111624 3.7 Faible

    Umbraco-CMS

    Umbraco CMS before 17.7.1 and 18.2.1 contains a missing authorization vulnerability that allows unauthenticated clients to connect to the backoffice PreviewHub and BackofficeHub SignalR hubs. Attackers can complete the SignalR negotiate step and receive refreshed notifications that disclose content keys, including unpublished content, and reveal editorial activity timing.

  45. CVE-2026-111622 6.8 Moyenne

    a466350665 smart-sso

    smart-sso through 2.1.1 contains an open redirect vulnerability in the /login handler that allows unauthenticated attackers to steal OAuth2 authorization codes by supplying an unvalidated redirectUri parameter. Attackers can send crafted /login links to users with active SSO sessions, then redeem stolen codes at /oauth2/access_token with any valid client credentials to obtain victims' access tokens.

  46. CVE-2026-111617 6.5 Moyenne

    MGdaasLab WHartTest

    WHartTest through 2.8.1 contains a missing authentication vulnerability that allows unauthenticated attackers to connect to the /ws/ui/web/ and /ws/ui/actuator/ WebSocket endpoints because UiAutomationConsumer accepts every connection. Attackers can register an actuator with a client-supplied id to intercept other users' test-case payloads, receive broadcast step results, and spoof results and screen frames into victim sessions.

  47. CVE-2026-111616 4.9 Moyenne

    Oqtane Framework

    Oqtane Framework through 10.2.6 contains an authorization bypass vulnerability in SettingController that allows a site administrator to read another co-tenant site's private settings. Attackers with the Administrators role on one site can request another site's entityId through their own alias to retrieve stored SMTPPassword and SMTPClientSecret values.

  48. CVE-2026-109461 5.9 Moyenne

    Microsoft Identity Client for NodeJS

    Origin validation error in Microsoft Authentication Library (MSAL) for Node.js allows an unauthorized attacker to disclose information over a network.

  49. CVE-2026-49868 9.8 Critique

    RIOT-OS RIOT

    RIOT is an open-source microcontroller operating system designed for Internet of Things devices and other embedded systems. Prior to 2026.07, _parse_reply() in sys/net/application_layer/dhcpv6/client.c omits minimum-length validation before passing IA_PD and IA_NA option bodies to sub-option parsers. A malicious or compromised DHCPv6 server can send a REPLY with an option length smaller than the required structure while MODULE_DHCPV6_CLIENT_IA_PD is enabled, causing _parse_ia_pd_option to read past the option body, size_t underflow, and unbounded out-of-bounds heap reads that can crash the device and may expose or corrupt memory. This issue is fixed in version 2026.07.

  50. CVE-2026-102692 9.1 Critique

    RIOT-OS RIOT

    RIOT is an open-source microcontroller operating system designed for Internet of Things devices and other embedded systems. Prior to 2026.07, nanocoap accepts oversized CoAP Block2 option values without enforcing the 20-bit maximum block number, allowing block offset multiplication to overflow on 32-bit targets. The affected master baseline is commit aeeae69e; the verified fix is later on master and remains unreleased. The oversized value passes from coap_get_blockopt() to coap_block_slicer_init(). A remote CoAP client can send a malformed Block2 option that produces an invalid slice range and can cause out-of-bounds memory access, denial of service, or memory corruption. This vulnerability is fixed in 2026.07.

  51. CVE-2026-68908 5.3 Moyenne

    mz-automation libiec61850

    libIEC61850 is an open-source implementation of IEC 61850 client and server protocols including MMS, GOOSE, Sampled Values, and R-Session. Prior to 1.6.2, Reporting_RCBWriteAccessHandler accepts an RptID longer than 129 bytes in the report-control-block write path. When GI later causes sendNextReportEntrySegment to serialize the report, the function wraps rptIdBuf[130] as a temporary MmsValue and passes it to MmsValue_setVisibleString. The setter's resize path frees the stack-backed buffer, allowing a remote client to trigger an invalid free and server crash by writing the oversized RptID and then generating a report. This issue is fixed in version 1.6.2.

  52. CVE-2026-68907 8.2 Élevée

    mz-automation libiec61850

    libIEC61850 is an open-source implementation of IEC 61850 client and server protocols including MMS, GOOSE, Sampled Values, and R-Session. Prior to 1.6.2, checkDataSetAccess reverses the source and destination arguments to StringUtils_copyStringToBuffer while constructing an MMS_ASSOCIATION_SPECIFIC dataset reference. The helper treats uninitialized dataSetRef[130] storage as a source string, causing strlen to read beyond the stack buffer, and then copies those bytes into the caller-owned listName buffer. A remote defineNamedVariableList request can crash servers that have dynamic dataset service enabled and a dataset access handler installed; the advisory also observes caller stack corruption but establishes denial of service as the primary impact. This issue is fixed in version 1.6.2.

  53. CVE-2026-68906 8.7 Élevée

    mz-automation libiec61850

    libIEC61850 is an open-source implementation of IEC 61850 client and server protocols including MMS, GOOSE, Sampled Values, and R-Session. Prior to 1.6.2, parseUserInformation in src/mms/iso_acse/acse.c checks len for zero before validating the negative bufPos returned by BerDecoder_decodeLength for a truncated long-form length. The continue path therefore swallows the error, reads buffer[-1], resets parsing to the beginning, and repeats the same malformed state. An unauthenticated remote peer can send a malformed AARE PDU during association to cause an infinite CPU loop under a normal allocator or a one-byte out-of-bounds read and process abort when the preceding byte is inaccessible. This issue is fixed in version 1.6.2.

  54. CVE-2026-57446 5.3 Moyenne

    mz-automation libiec61850

    libIEC61850 is an open-source implementation of IEC 61850 client and server protocols including MMS, GOOSE, Sampled Values, and R-Session. From 1.6.1 until 1.6.2, a successful MMS ObtainFile operation closes MmsObtainFileTask::fileHandle but leaves the stale non-NULL value in a task marked MMS_FILE_UPLOAD_STATE_NOT_USED. If a later ObtainFile request reuses that task and destination creation fails, MMS_FILE_UPLOAD_STATE_SEND_OBTAIN_FILE_ERROR_DESTINATION cleanup closes the stale FILE object again. Servers with the MMS file service and ObtainFile enabled can therefore be crashed remotely by the success-then-failure request sequence. This issue is fixed in version 1.6.2.

  55. CVE-2026-55836 6.8 Moyenne

    open-zaak

    Open Zaak is an open-source data and services layer for Dutch case-oriented working. From 1.13.0 until 1.27.2, 1.28.2, and 1.29.0, _import_document_row in src/openzaak/components/documenten/tasks.py uses the attacker-controlled bestandspad CSV field without ensuring that the resolved path remains under IMPORT_DOCUMENTEN_BASE_DIR. A client with an all-authorizations application token can supply an absolute or parent-traversal path to a file readable by the Celery service account. The worker copies that file into document storage, allowing its contents to be retrieved through the normal document download API and exposing server secrets or other private files. This issue is fixed in versions 1.27.2, 1.28.2, and 1.29.0.

  56. CVE-2026-48741 7.5 Élevée

    mz-automation libiec61850

    libIEC61850 is an open-source implementation of IEC 61850 client and server protocols including MMS, GOOSE, Sampled Values, and R-Session. Prior to 1.6.2, alternateArrayAccess in src/mms/iso_mms/server/mms_read_service.c does not validate an alternate-access sub-array range when numberOfElements is greater than zero. An out-of-range MmsValue_getElement call returns NULL, and the sub-array branch passes that value to MmsValue_isDeletable, causing a NULL pointer dereference. A remote MMS Read request can therefore crash servers whose data models contain arrays, while the single-element error path handles the same invalid index safely. This issue is fixed in version 1.6.2.

  57. CVE-2026-48503 7.5 Élevée

    mz-automation libiec61850

    libIEC61850 is an open-source implementation of IEC 61850 client and server protocols including MMS, GOOSE, Sampled Values, and R-Session. Prior to 1.6.2, the MMS file service in src/mms/iso_mms/server/mms_file_service.c accepts an out-of-range initialPosition in fileOpen and stores it in the file-read state. The fileRead path then subtracts the larger read position from the file size using unsigned arithmetic and fails to limit the response to the actual FileSystem_readFile return value. The oversized response can include stale bytes from the reused connection buffer after the requested file contents, disclosing prior response data and protocol fragments to an unauthenticated remote client when the file service is enabled. This issue is fixed in version 1.6.2.

  58. CVE-2026-111728 5.4 Moyenne

    5rahim Seanime

    Seanime v3.7.0 to v3.10.3 contains an argument injection vulnerability reachable through improper authorization, allowing unauthenticated LAN or Tailnet clients to bypass the privileged mediastream settings guard by setting transcodeHwAccel to custom and supplying attacker-controlled decodeFlags and encodeFlags via the transcodeHwAccelCustomSettings field, which are appended verbatim to the ffmpeg argument vector. Attackers can trigger execution via the mediastream request and HLS playlist endpoints to achieve arbitrary file write and file truncation on the host as the user running seanime, with the injected input options also enabling potential arbitrary local file read and outbound data exfiltration through ffmpeg's protocol handlers.

  59. CVE-2026-97470 6.4 Moyenne

    themefic Tourfic – AI Powered Travel Booking, Hotel Booking & Car Rental

    The Tourfic – AI Powered Travel Booking, Hotel Booking & Car Rental plugin for WordPress is vulnerable to Stored Cross-Site Scripting via Comment Content via unfiltered_html Grant to Customer Role in all versions up to, and including, 3.0.1 due to insufficient input sanitization and output escaping. This makes it possible for authenticated attackers, with subscriber-level access and above, to inject arbitrary web scripts in pages that will execute whenever a user accesses an injected page. This requires WooCommerce to be active, but since the customer role is freely self-registerable without admin approval, exploitation is effectively possible by any unauthenticated visitor who registers a WooCommerce account. While the plugin's own listing-page templates sanitize output with wp_kses_post(), the stored payload is returned unsanitized via the unauthenticated GET /wp-json/wp/v2/comments REST endpoint (content.rendered) and post/comment RSS feeds, exposing any headless client, feed reader, or third-party widget that renders comment content.

  60. CVE-2026-110804 4.3 Moyenne

    Devolutions Server · Devolutions devolutions server

    Improperly controlled modification of JIT elevation settings in PAM checkout in Devolutions Server 2026.3.8.0 and earlier allows an authenticated user with PAM checkout access to delete arbitrary Active Directory groups or retain elevated privileges via client-supplied JIT elevation settings in a checkout update request.

  61. CVE-2026-105958 7.5 Élevée

    Red Hat Enterprise Linux 6 · Red Hat Enterprise Linux 7 · Red Hat Enterprise Linux 8

    A flaw was found in spice-server. Multiple migration data handlers use client-supplied offsets from SPICE_MSGC_MIGRATE_DATA messages without bounds validation, allowing a malicious authenticated SPICE client during VM live migration to read arbitrary heap memory or trigger heap corruption via out-of-bounds read and write operations on the message buffer.

  62. CVE-2026-105959 7.5 Élevée

    Red Hat Enterprise Linux 6 · Red Hat Enterprise Linux 7 · Red Hat Enterprise Linux 8

    A flaw was found in spice-server. The GLZ encoder dictionary restore path truncates a size_t allocation size to int when calling the malloc callback, allowing a malicious authenticated SPICE client during VM live migration to trigger a massive heap buffer overflow by sending a crafted max_encoders value in migration data.

  63. CVE-2026-14452 8.6 Élevée

    Checkmk GmbH Checkmk

    Improper validation of the reset_url parameter of the Setup in-page search in Checkmk <2.5.0p15, <2.4.0p38, <2.3.0p51 and 2.2.0 (EOL) allows an attacker to execute arbitrary script in the session of an authenticated user with Setup access, or to redirect that user to an arbitrary external site, by tricking the user into opening a crafted URL and clicking the Reset button or submitting an empty search.

  64. CVE-2026-19541 6.5 Moyenne

    Mattermost

    Mattermost versions 11.9.x <= 11.9.1, 11.8.x <= 11.8.5, 11.7.x <= 11.7.10, 11.10.x <= 11.10.1 fail to require re-entry of the stored client secret when validating an outgoing OAuth connection, which allows an authenticated user with permission to manage outgoing OAuth connections to obtain the secret via pointing the token URL at an attacker-controlled endpoint and triggering validation.. Mattermost Advisory ID: MMSA-2026-00761

  65. CVE-2026-111069 6.4 Moyenne

    paoltaia GeoDirectory – WP Business Directory Plugin and Classified Listings Directory

    The GeoDirectory – WP Business Directory Plugin and Classified Listings Directory plugin for WordPress is vulnerable to Stored Cross-Site Scripting via the 'twitterusername' parameter in all versions up to, and including, 2.8.192 due to insufficient input sanitization and output escaping. This makes it possible for authenticated attackers, with subscriber-level access and above, to inject arbitrary web scripts in pages that will execute whenever a user accesses an injected page. The client-side HTML pattern attribute enforcing alphanumeric-only input on the twitterusername field is not validated server-side, allowing a listing owner with Subscriber-level access to submit an attribute-breakout payload that bypasses sanitization and is stored directly to the custom field table.

  66. CVE-2026-110135 7.4 Élevée

    Unknown OTP Login & Register Woocommerce

    The OTP Login & Register Woocommerce WordPress plugin before 2.7.3 does not adequately limit one-time-password verification attempts, because the attempt limit is tracked against a value the client itself supplies and can change freely, allowing an unauthenticated attacker to bypass the limit and brute-force the short numeric login code to take over an account without its password.

  67. CVE-2026-110015 8.8 Élevée

    Unknown Salon Booking System

    The Salon Booking System WordPress plugin before 10.31.9 does not correctly verify the authenticity of REST API requests and discards WordPress core's nonce check for cookie-authenticated sessions across the whole REST API, allowing unauthenticated attackers who trick a logged-in administrator into clicking a crafted link to forge requests in that administrator's browser, up to creating a new administrator account.

  68. CVE-2026-104730 9.8 Critique

    sunshinephotocart Sunshine Photo Cart – Client Photo Gallery & Photo Proofing for Photographers

    The Sunshine Photo Cart – Client Photo Gallery & Photo Proofing for Photographers plugin for WordPress is vulnerable to PHP Object Injection in all versions up to, and including, 3.7.4 via deserialization of untrusted input . This makes it possible for unauthenticated attackers to inject a PHP Object. No known POP chain is present in the vulnerable software, which means this vulnerability has no impact unless another plugin or theme containing a POP chain is installed on the site. If a POP chain is present via an additional plugin or theme installed on the target system, it may allow the attacker to perform actions like delete arbitrary files, retrieve sensitive data, or execute code depending on the POP chain present. If a POP chain is present via an additional plugin or theme installed on the target system, the attacker can store the malicious payload unauthenticated via the public favorites page and trigger deserialization by viewing any image in a non-viewable or private gallery.

  69. CVE-2026-89051 7.2 Élevée

    johndarrel Hide My WP Ghost – Security & Firewall

    The WP Ghost (Hide My WP Ghost) – Security & Firewall plugin for WordPress is vulnerable to Stored Cross-Site Scripting via the 'replaceHeadersAndTags' function in versions up to, and including, 7.0.12 due to insufficient input sanitization and output escaping. This makes it possible for unauthenticated attackers to inject arbitrary web scripts in pages that will execute whenever a user clicks on the injected link.

  70. CVE-2026-111122 7.2 Élevée

    gpriday SiteOrigin CSS

    The SiteOrigin CSS plugin for WordPress is vulnerable to Stored DOM-Based Cross-Site Scripting via the 'href' parameter in all versions up to, and including, 1.6.7 due to insufficient input sanitization and output escaping. This makes it possible for unauthenticated attackers to inject arbitrary web scripts in pages that will execute whenever a user accesses an injected page. Successful exploitation requires the malicious comment to be approved and an administrator to subsequently open the Visual Inspector and click the attacker's anchor link, both of which are plausible interactions during routine comment triage.

  71. CVE-2026-109447 8.1 Élevée

    cozmoslabs User Profile Builder – Beautiful User Registration Forms, User Profiles & User Role Editor

    The User Profile Builder – Beautiful User Registration Forms, User Profiles & User Role Editor plugin for WordPress is vulnerable to Privilege Escalation in all versions up to, and including, 4.0.3 via the `role` parameter handled by the `wppb_add_meta_to_user_on_activation()` function. The vulnerability exists because the activation handler passes the attacker-supplied role value directly to `WP_User::set_role()` with no allow-list validation, and the 4.0.3 reserved-name check exempts pre-existing form fields from the block, allowing a legacy field with meta-name `role` to persist and bridge attacker-controlled request input into the serialized signup record stored in `wp_signups`. This makes it possible for unauthenticated attackers to gain a fully activated WordPress administrator account by submitting `role=administrator` through the public `[wppb-register]` shortcode form and subsequently clicking the confirmation activation link. Exploitation requires the site to have a legacy form field with meta-name `role` carried over from before the 4.0.3 upgrade, email confirmation enabled, and open user registration — all stable site-state conditions once established.

  72. CVE-2026-109400 7.5 Élevée

    InstaWP Connect – 1-click WP Staging & Migration

    The InstaWP Connect – 1-click WP Staging & Migration plugin for WordPress is vulnerable to Privilege Escalation in all versions up to, and including, 0.1.4.1 This is due to the plugin registering its admin page via `add_management_page()` using a role name returned by `InstaWP_Setting::get_allowed_role()` as the required capability — which WordPress resolves as a real capability match for any user holding that role — combined with `generate_section_field()` rendering the master API key from the `instawp_api_options` option directly as the `value` attribute of a plain-text HTML input with no additional `manage_options` or equivalent capability check. This makes it possible for authenticated attackers with Subscriber-level access and above to read the plaintext master API key, use its SHA-256-derived secret as a bearer token accepted by the plugin's REST API `validate_api_request()`, and call `POST /wp-json/instawp-connect/v2/auto-login` to mint a 30-second single-use login URL that sets an administrator authentication cookie — fully escalating their privileges. Exploitation requires the site to be actively connected to InstaWP and an administrator to have added the attacker's role to the plugin's Sync Tab Access list (`instawp_sync_tab_roles`); notably, the settings tab suppression logic only hides the tab when `administrator` is absent from the list entirely, meaning a mixed configuration retaining administrator alongside a lower role still serves the full settings page — including the API key — to the lower-privileged user.

  73. CVE-2026-109191 4.3 Moyenne

    extendthemes Image Hub – Free Images from Unsplash, Pixabay, Pexels, Openverse & Giphy

    The Image Hub – Free Images from Unsplash, Pixabay, Pexels, Openverse & Giphy plugin for WordPress is vulnerable to Sensitive Information Exposure in all versions up to, and including, 1.2.0 via the 'image_hub_api_keys_* (6 options)' parameter. This makes it possible for authenticated attackers, with subscriber-level access and above, to extract the plaintext API credentials for six third-party services — Giphy, Openverse (client ID and secret), Pexels, Pixabay, and Unsplash — emitted verbatim into every admin page's HTML via wp_localize_script. The leaked page HTML also contains a valid image_hub_nonce AJAX nonce, which — combined with the absence of any capability check in the image_hub_update_settings AJAX handler — allows the same Subscriber-level attacker to overwrite all six stored API credential options via update_option().

  74. CVE-2026-103481 6.1 Moyenne

    mpol Gwolle Guestbook

    The Gwolle Guestbook plugin for WordPress is vulnerable to Reflected Cross-Site Scripting via the 'permalink' parameter in all versions up to, and including, 5.1.1 due to insufficient input sanitization and output escaping. This makes it possible for unauthenticated attackers to inject arbitrary web scripts in pages that execute if they can successfully trick a user into performing an action such as clicking on a link.

  75. CVE-2026-111451 7.8 Élevée

    Wheels CLI

    Wheels CLI 4.0.0 through 4.1.1 insufficiently validates or escapes caller-supplied values and package contents. Deploy release/version/rollback values can execute shell commands on deploy hosts as the SSH user and on the CI runner during builds. Package removal and destroy paths can delete unintended files; package extraction can overwrite other vendor files; generator names can write outside the project or inject executable CFML into generated code. Inputs may originate from CI variables, branch names, packages or an AI agent using the CLI's MCP tools. Also relevant are unencoded display properties in generated scaffold show views since 4.1.0 and in generated admin views, permitting stored XSS, and stale process-ID handling in RustCFML server commands that can terminate an unrelated process. This is fixed in 4.1.2; existing affected generated views also require the documented edit.

  76. CVE-2026-111449 5.9 Moyenne

    Wheels

    Wheels 3.0.0 through 4.1.1 can disclose detailed exceptions when allowIPBasedDebugAccess=true (disabled by default). An allowed administrator request changes error/debug-display settings in shared application scope. If it overlaps with an unauthorized client's request that raises an error, that client can receive full exception details, including tag context or a serialized exception, and any sensitive data they contain. This race does not bypass the production gate protecting developer tools. This is fixed in 4.1.2.

  77. CVE-2026-111448 3.7 Faible

    Wheels

    Wheels 4.0.0 through 4.1.2 uses case-insensitive comparisons in path-containment checks on case-sensitive filesystems. Canonical-path checks for zip entries, package mappings, development assets/docs, guide images and image tags can accept paths through an existing symlink to a case-distinct sibling directory, allowing access outside the intended root, including limited file disclosure. Separately, the CLI database-dump output check can write outside the application root to an existing writable case-distinct sibling directory; that action requires the CLI POST, loopback and password gate. This is fixed in 4.2.0.

  78. CVE-2026-111446 8.6 Élevée

    Wheels

    Wheels 4.0.0 through 4.1.2 exposes password-less development tools under /wheels/* to remote clients when enablePublicComponent=true, as it is by default in development. Before 4.1.2 these tools have no local-access restriction; in 4.1.2 the restriction trusts client-supplied headers without checking the connecting peer. Remote clients can obtain application settings, environment/datasource details and routes, and start test runs. Where application tests use the primary datasource, those runs can write to the real application database. The separately protected migrator and CLI endpoints are not affected. This is fixed in 4.2.0.

  79. CVE-2026-111443 6.5 Moyenne

    Wheels

    Wheels 4.0.6 through 4.1.2 lets an unauthenticated client switch a request into a separate test application context, including in production, by supplying an X-Wheels-Test-Context header or WHEELS_TEST_CONTEXT cookie with any value, or including a test-runner path anywhere in the URL, even the query string. The alternate application uses the same configuration, environment, datasource and debug settings: this is not an authentication bypass or data/debug escalation. It can bypass maintenance mode enabled only at runtime, provide a second set of in-memory rate-limit counters, and trigger additional application-startup side effects and memory use. This is fixed in 4.2.0.

  80. CVE-2026-111442 5.3 Moyenne

    BackdropCMS Backdrop

    Backdrop CMS Core 1.35.x before 1.35.2 and 1.34.x before 1.34.6 does not sufficiently validate uploaded SVG images to ensure they lack potentially dangerous SVG tags. SVG images can contain clickable links or executable scripting. By uploading a crafted SVG, it is possible to execute scripting in the browser when an SVG image is viewed. This issue is mitigated by the attacker needing to be able to upload SVG images, and that Backdrop embeds all uploaded SVG images within <img> tags, which prevents script from executing. The SVG must be viewed directly by its URL in order to run any embedded scripting.

  81. CVE-2026-111382 8.8 Élevée

    FreeRDP · Freerdp freerdp

    FreeRDP before 3.32.0 contains a client-side heap out-of-bounds write vulnerability in general_LumaToYUV444 and general_ChromaV2ToYUV444 that allows malicious RDP servers to corrupt client memory. Attackers can send RDPGFX AVC444 or AVC444v2 frames with odd-height region rects ending at surface height to write past U/V planes, enabling heap corruption and potential code execution.

  82. CVE-2026-111380 8.8 Élevée

    FreeRDP · Freerdp freerdp

    FreeRDP from 3.29.0 before 3.32.0 contains an integer overflow in the Apple audin backends (audin_mac.m, audin_ios.m), which compute the audio queue buffer size as FramesPerPacket * 2 * nChannels in UINT32 arithmetic, bypassing the CVE-2026-91962 fix. A malicious RDP server can send FramesPerPacket of 2^30 in the audin Open PDU to wrap the multiplication, causing undersized AudioQueueAllocateBuffer allocations and a heap out-of-bounds write in macOS and iOS clients.

  83. CVE-2026-111377 6.5 Moyenne

    FreeRDP · Freerdp freerdp

    FreeRDP from 2.0.0 before 3.32.0 contains an unbounded reassembly vulnerability in WTSProcessChannelData that fails to enforce cumulative fragment limits. Authenticated clients can start a static virtual channel message, withhold the LAST flag, and send endless fragments to exhaust server memory.

  84. CVE-2026-111376 5.0 Moyenne

    FreeRDP · Freerdp freerdp

    FreeRDP from 3.30.0 before 3.32.0 contains a heap out-of-bounds read vulnerability in the Windows client's CliprdrStream_New that allows malicious RDP servers to trigger invalid memory reads via short clipboard responses. Attackers can answer a FILECONTENTS_SIZE request with a cbRequested of fewer than eight bytes when a user pastes a remote file, crashing the client or corrupting the stream size.

  85. CVE-2026-111375 9.1 Critique

    FreeRDP · Freerdp freerdp

    FreeRDP from 3.0.0 before 3.32.0 contains an out-of-bounds read vulnerability in the Windows RAIL client that allows malicious RDP servers to trigger heap over-reads by sending non-NUL-terminated window titles. Attackers can send a full-length WINDOW_ORDER_FIELD_TITLE update that wf_rail_window_common passes to SetWindowTextW, corrupting titles or crashing the client.

  86. CVE-2026-111374 5.3 Moyenne

    FreeRDP · Freerdp freerdp

    FreeRDP from 3.30.0 before 3.32.0 contains a resource exhaustion vulnerability in the rdpecam client that lets malicious RDP servers request unadvertised camera media types. Attackers can send a StartStreamsRequest with oversized dimensions to force large pending-sample allocations and reconfigure the local camera through the HAL.

  87. CVE-2026-111373 7.5 Élevée

    FreeRDP · Freerdp freerdp

    FreeRDP before 3.32.0 contains a NULL pointer dereference vulnerability in the websocket transport that allows a malicious gateway to crash clients by sending zero-length pings. When the payload length is zero, websocket_reply_pong passes a nullptr stream to websocket_reply_close, triggering an assertion failure that terminates the client process and causes denial of service.

  88. CVE-2026-100159 4.8 Moyenne

    ays-pro Popup Box – Create Countdown, Coupon, Video, Contact Form Popups

    The Popup Box plugin for WordPress is vulnerable to Arbitrary Shortcode Execution in versions 6.4.1 to 6.4.2. This is due to the ays_popup_get_user_ip() function reading the client-controlled Client-IP / X-Forwarded-For / Forwarded / X-Forwarded / Forwarded-For request headers, sanitizing them only with sanitize_text_field() (which does not strip the bracket, quote, space, or '=' characters used by shortcode syntax), substituting the value into a popup body wherever the %%user_ip_address%% token appears (a raw str_replace with no escaping in Ays_Pb_Data::replace_message_variables()), and then passing the resulting body to do_shortcode() in Ays_Pb_Public::ays_autoembed() at output on the wp_footer hook. This makes it possible for unauthenticated attackers to execute arbitrary registered shortcodes with attacker-supplied attributes on the server whenever an administrator has added the %%user_ip_address%% token to a popup's Custom Content body, which can be leveraged to read arbitrary data or perform other actions exposed by shortcodes installed on the site.

  89. CVE-2026-92629 7.3 Élevée

    Lenovo Baiying PC Client

    A potential improper authorization vulnerability was reported in Lenovo Baiying PC Client, distributed exclusively in the Chinese market, that could allow a local authenticated user to execute arbitrary code with elevated privileges.

  90. CVE-2026-94135 4.3 Moyenne

    Apache Software Foundation Apache Airflow SFTP provider

    The Apache Airflow SFTP provider's deferrable code paths (`SFTPTrigger`, `SFTPHookAsync`, `SFTPClientPool`) disable SSH host-key verification by default unless the connection explicitly sets `no_host_key_check` to false. Deployments that run SFTP sensors or operators in deferrable mode are affected. An attacker in a machine-in-the-middle position between the triggerer and the SFTP server can impersonate the server and capture the credentials or data exchanged. Users of apache-airflow-providers-sftp are recommended to upgrade to apache-airflow-providers-sftp 7.0.0 or later. This is a separate defect from CVE-2026-94134 (the synchronous SSH hook default, which also affects the SFTP hook); users who applied the fix for one also need this release for the other - apache-airflow-providers-sftp 7.0.0 fixes both.

  91. CVE-2026-111313 6.5 Moyenne

    DaKheera47 JobOps

    JobOps through 0.13.1 contains a server-side request forgery vulnerability in the POST /api/manual-jobs/fetch handler that allows authenticated users, or unauthenticated clients in demo mode, to make the server fetch internal destinations. Attackers can supply loopback, private, or cloud metadata addresses in the url field to read returned page content, disclosing instance metadata and internal service data.

  92. CVE-2026-111280 7.4 Élevée

    davenardella snap7

    snap7 through 1.4.3 contains an out-of-bounds write vulnerability in TSnap7MicroClient::opReadArea() that allows malicious S7 peers to overflow client buffers via an unchecked DataLength field. Rogue or impersonated PLCs can return crafted ReadVar responses causing memcpy of up to 65535 bytes into caller buffers, crashing SCADA, HMI, or gateway processes and corrupting memory.

  93. CVE-2026-111279 7.5 Élevée

    dragonflydb dragonfly

    Dragonfly through 2.0.2 contains an unbounded memory allocation vulnerability that allows unauthenticated clients to exhaust memory by passing a negative COUNT to HRANDFIELD, ZRANDMEMBER or SRANDMEMBER. Attackers can send a single command against a one-element key with a huge negative COUNT to trigger the OOM killer, terminating the instance for all clients.

  94. CVE-2026-111275 6.1 Moyenne

    NIFTI-Imaging nifti_clib

    nifti_clib through 3.0.1 contains a stack-based out-of-bounds read and write vulnerability in nifti_convert_n2hdr2nim() because the NIfTI-2 dim[0] value is never range-checked. Attackers can supply a crafted 540-byte NIfTI-2 header read via nifti_image_read() to crash applications or corrupt stack memory.

  95. CVE-2026-94621 6.3 Moyenne

    aliasvault · aliasvault ghcr.io/aliasvault/aliasvault · aliasvault ghcr.io/aliasvault/api · aliasvault ghcr.io/lanedirt/aliasvault-api

    AliasVault is a privacy-first password manager with built-in email aliasing. Prior to 0.30.3, the AliasVault server API favicon extractor validates a user-supplied hostname with a first DNS lookup but allows the HTTP client to perform a second DNS lookup when connecting. An authenticated low-privileged user operating an attacker-controlled authoritative DNS server can return a public address during validation and a private address during connection, causing a publicly reachable self-hosted server to issue HTTP or HTTPS requests to internal hosts. A valid image response is returned in Base64 form, while other responses can expose limited timing and error information for internal-network reconnaissance. Exploitation requires a publicly reachable self-hosted instance with public registration enabled, an authenticated AliasVault account, an attacker-controlled authoritative DNS server, and a resolver that permits rebinding; default checks limit requests to ports 80 and 443. The browser extensions, web application, and mobile applications are not affected. This issue is fixed in version 0.30.3.

  96. CVE-2026-87029 7.5 Élevée

    undici · Nodejs undici

    undici is an HTTP client library for Node.js. In versions 7.15.0 through 7.30.0 and 8.0.0 through 8.12.1, its decompress interceptor forwards decompressed response data to the downstream handler from zlib stream event callbacks without catching exceptions those callbacks may throw, and while undici's engine converts a synchronous throw from a downstream handler into a per-request failure, that protection is absent when the throw happens asynchronously from a zlib callback. As a result, a downstream handler that throws by design, such as the built-in dump interceptor which throws when a response exceeds its configured maximum size, escapes as an uncaught exception that terminates the Node.js process. A malicious or compromised HTTP server can crash any client that composes the decompress interceptor with such a handler by returning a small compressed body that decompresses to more than the configured limit, so only the decompressed size matters and the bytes on the wire stay tiny. The issue is fixed in undici 7.30.1 and 8.12.2, and users should upgrade to one of those versions or later. As a workaround, avoid composing the decompress interceptor with a response handler that throws during response processing until upgrading.

  97. CVE-2026-87018 7.5 Élevée

    undici · Nodejs undici

    undici is an HTTP client library for Node.js. In versions 4.2.1 through 6.29.0, 7.0.0 through 7.30.0, and 8.0.0 through 8.12.1, the client process crashes when a TLS 1.2 server renegotiates after the connection is established and presents a different, CA-trusted certificate whose hostname does not match the requested host. Node.js re-runs hostname verification on the renegotiated handshake and destroys the socket with a certificate-altname error, but undici's established-socket error handler asserted that this error code could never occur, so the peer-controlled error became an uncaught assertion that terminates the Node.js process, affecting both the HTTP/1.1 and HTTP/2 paths. A malicious or compromised HTTPS endpoint reached through undici or the global fetch() can crash the client with a single server-initiated renegotiation, taking down all in-flight requests and background work in the process, and exploitation needs no CA compromise because the attacker controls the endpoint and can present two ordinary valid certificates. Only TLS 1.2 is affected, since TLS 1.3 does not support renegotiation. The issue is fixed in undici 6.29.1, 7.30.1, and 8.12.2, and users should upgrade to one of those versions or later. As a workaround, restrict outbound connections to TLS 1.3 so renegotiation cannot occur, when the target endpoints support it.

  98. CVE-2026-57496 9.6 Critique

    Labs64 NetLicensing-MCP

    NetLicensing MCP Server is a natural-language interface that enables agentic applications to manage the software-licensing lifecycle in Labs64 NetLicensing. Prior to 0.1.8, netlicensing_get_product passes the caller-controlled product_number to src/netlicensing_mcp/tools/products.py, which interpolates it into a REST path consumed by nl_get() in src/netlicensing_mcp/client.py without path-segment validation. A traversal segment can be normalized from the intended product path to the /token endpoint, after which the response is serialized through _wrap_json instead of _wrap_json_token_read, bypassing redact_token_read() protections for APIKEY number and SHOP shopURL fields. An authenticated MCP client with a non-administrative API key can retrieve plaintext token values, and a recovered administrative token can provide broad read, write, and delete access to the NetLicensing account. This issue is fixed in version 0.1.8.

  99. CVE-2026-36436 7.5 Élevée

    An issue in Trisolutions BRPerfX version 1.4.4.0 exposes LDAP credentials in plaintext within the administrative configuration interface. An authenticated administrator can inspect the client-side code to retrieve the credential.

  100. CVE-2026-111230 4.9 Moyenne

    Netgate pfSense CE · Pfsense pfsense

    Netgate pfSense CE through 2.9.0 contains a local file read vulnerability that allows users with the Dynamic DNS client privilege to read files as root by setting a file:// custom Update URL in services_dyndns_edit.php. Attackers can set a non-matching Result Match string so _checkStatus() logs file contents such as /conf/config.xml or /etc/master.passwd to the system log.

  101. CVE-2026-111228 5.5 Moyenne

    Netgate pfSense CE · Pfsense pfsense

    Netgate pfSense CE from 2.7.1 through 2.9.0 contains a CRLF injection vulnerability in services_unbound_advanced.php that allows authenticated users with DNS Resolver: Advanced privilege to inject Unbound directives. Attackers can embed CR/LF characters in the unvalidated sock_queue_timeout parameter, which is written raw into unbound.conf, to override local records and forwarding for all clients.

  102. CVE-2026-111227 5.5 Moyenne

    Netgate pfSense CE · Pfsense pfsense

    Netgate pfSense CE through 2.9.0 contains a CRLF injection vulnerability in services_unbound_acls.php that allows authenticated users with the DNS Resolver: Access Lists privilege to inject Unbound configuration directives. Attackers can embed CR/LF characters in the aclname or aclaction fields written to access_lists.conf to add local-data, local-zone, or forward-zone directives and control DNS answers served to LAN clients.

  103. CVE-2026-111225 8.8 Élevée

    Netgate pfSense CE · Pfsense pfsense

    Netgate pfSense CE 2.5.0 through 2.9.0 contains an OS command injection vulnerability in openvpn.learn-address.sh that allows authenticated VPN clients to execute commands as the unbound user. Attackers can embed shell metacharacters in the certificate common name or username, which is interpolated unquoted into su -c strings, when DNS Resolver client registration is enabled.

  104. CVE-2026-111219 7.2 Élevée

    Netgate pfSense CE · Pfsense pfsense

    Netgate pfSense CE through 2.9.0 contains a newline injection vulnerability in services_pppoe_edit.php that allows privileged users to inject mpd.secret entries via unvalidated username POST fields. Attackers with the PPPoE Server Edit privilege can add a line whose secret begins with '!', making mpd5 run commands as root when a PPPoE client authenticates.

  105. CVE-2026-111171 5.4 Moyenne

    ethereum web3.py

    Web3.py 7.15.0 through 7.16.0 and 8.0.0b2 through 8.0.0 contains a server-side request forgery vulnerability because _check_ip_blocked() does not unwrap IPv4-mapped IPv6 addresses in CCIP Read URL validation. Malicious contracts can return OffchainLookup URLs with IPv4-mapped IPv6 hosts or AAAA records, making clients with CCIP Read enabled send requests to loopback, private, or link-local services.

  106. CVE-2026-110361 6.5 Moyenne

    ClikIT Infinite Uploads

    Insertion of Sensitive Information Into Sent Data vulnerability in ClikIT Infinite Uploads infinite-uploads allows Retrieve Embedded Sensitive Data.This issue affects Infinite Uploads: from n/a through 3.4.1.

  107. CVE-2026-109932 7.5 Élevée

    undici · Nodejs undici

    undici is an HTTP client library for Node.js. In versions 5.15.0 through 6.29.0, 7.0.0 through 7.30.0, and 8.0.0 through 8.12.1, its WebSocket client, including the bundled WebSocket exposed by Node.js, can be driven to exhaust process memory by a malicious or compromised WebSocket server. For every PING frame received, the client automatically writes a PONG response and ignores the backpressure signal from the socket write, so a server that streams small PING frames while never reading the client's PONGs makes the outbound write queue and heap grow without bound until the Node.js process runs out of memory and terminates. The attack is remote and unauthenticated, needs only a normal WebSocket connection, and is asymmetric, since small inbound PINGs produce a much larger retained outbound queue, with availability the only impact. The issue is fixed in undici 6.29.1, 7.30.1, and 8.12.2, which pause WebSocket frame parsing when the automatic PONG response hits socket backpressure and resume on drain, bounding the outbound queue, and users should upgrade to one of those versions or later. There is no workaround.

  108. CVE-2026-109924 7.5 Élevée

    undici · Nodejs undici

    undici is an HTTP client library for Node.js. In versions 6.5.0 through 6.29.0, 7.0.0 through 7.30.0, and 8.0.0 through 8.12.1, its EventSource client can be driven to exhaust process memory by a malicious or compromised Server-Sent Events peer. By sending millions of tiny completed data fields with no blank event separator, or a single unfinished line split across many chunks, the parser retains a growing amount of per-field and per-chunk state that the byte-based maxEventSize and maxResponseSize limits do not bound, and because maxResponseSize is enforced on the compressed body before decompression, a small compressed response can expand into enough retained memory to terminate the process. An application only needs to maintain an EventSource connection to an untrusted peer, which controls only the response, and availability is the only impact. The issue is fixed in undici 6.29.1, 7.30.1, and 8.12.2, which bound the retained parser state independently of the byte limits, and users should upgrade to one of those versions or later. There is no workaround.

  109. CVE-2026-109869 7.5 Élevée

    undici · Nodejs undici

    undici is an HTTP client library for Node.js. In versions 6.23.0 through 6.29.0, 7.18.0 through 7.30.0, and 8.0.0 through 8.12.1, its fetch() implementation crashes the Node.js process when a malicious server returns a response with more than five content codings and closes the connection before sending the body. undici rejects such responses because of a content-encoding limit, but it created the internal response stream before that check and returned without attaching an error listener or aborting the request, so the early close surfaces a content-length mismatch on the orphaned stream, which Node.js treats as an unhandled error and terminates the process. A remote, unauthenticated server can trigger this with a single response, even when the application correctly catches the fetch() rejection, resulting in a denial of service. The issue is fixed in undici 6.29.1, 7.30.1, and 8.12.2, and users should upgrade to one of those versions or later. There is no workaround.

  110. CVE-2026-109841 7.5 Élevée

    undici · Nodejs undici

    undici is an HTTP client library for Node.js. In versions 7.15.0 through 7.30.0 and 8.0.0 through 8.12.1, its fetch() implementation verifies Subresource Integrity metadata by matching the algorithm token against a lowercase-only set, even though the SRI specification defines the token by its ASCII-lowercase form, so a valid but non-lowercase token such as an uppercase or mixed-case SHA256 is treated as unknown and skipped. When no recognized algorithm remains, undici treats the integrity list as empty and accepts the response, so a server returning a body that does not match the pinned digest passes the check. An attacker who controls only the response body can therefore deliver attacker-controlled bytes to an application that pinned a valid uppercase or mixed-case SRI token, with no integrity error, affecting response integrity only. The issue is fixed in undici 7.30.1 and 8.12.2, which normalize the algorithm token to ASCII lowercase before validation, and users should upgrade to one of those versions or later. As a workaround, use lowercase algorithm tokens such as sha256, sha384, or sha512 in integrity metadata.

  111. CVE-2026-109829 9.1 Critique

    undici · Nodejs undici

    undici is an HTTP client library for Node.js. In versions 7.0.0 through 7.30.0 and 8.0.0 through 8.12.1, when an originless Agent composes the redirect and cache interceptors in that order with a shared cache store, the cache builds its key from the original request URL before the redirect handler changes the origin, so a response produced after a cross-origin redirect is stored under the original key. An attacker who controls the target of a redirect from a trusted URL can have their response cached under that trusted URL and served to later callers, allowing persistent cache poisoning, and the same key confusion lets a private, non-stored redirect response to a user-specific resource be stored under the original URL and served to a later unauthenticated caller, disclosing private data. Applications are affected only when an originless Agent, a shared cache store, and the redirect-then-cache composition are all used. The issue is fixed in undici 7.30.1 and 8.12.2, which verify the producing origin still matches the cache key and preserve each redirect hop's cache-control policy before storing, and users should upgrade to one of those versions or later. As a workaround, compose the cache interceptor before the redirect interceptor, use per-origin dispatchers and cache stores, or disable shared caching on redirect-capable requests.

  112. CVE-2026-109383 9.1 Critique

    undici · Nodejs undici

    undici is an HTTP client library for Node.js. In versions 4.0.0 through 6.29.0, 7.0.0 through 7.30.0, and 8.0.0 through 8.12.1, its Agent dispatcher passes its constructor options through a JSON-based deep clone before forwarding them to each per-origin pool, and because JSON cannot represent functions, a caller-supplied tls option containing a checkServerIdentity callback is silently dropped before it reaches the TLS layer. As a result, a TLS peer whose certificate a custom checkServerIdentity was written to reject, but which passes Node's default hostname and chain checks, is silently accepted. Because Agent is the default dispatcher behind request, fetch, and setGlobalDispatcher, this affects the path most applications use, and EnvHttpProxyAgent is affected on its direct and no-proxy routes as well, while only applications that pass a function-valued tls option such as a custom checkServerIdentity for certificate or public-key pinning are at risk. The issue is fixed in undici 6.29.1, 7.30.1, and 8.12.2, which preserve the tls option outside the JSON clone so custom verification callbacks are forwarded unchanged, and users should upgrade to one of those versions or later. As a workaround, pass the TLS options through connect instead of tls, or for proxied destinations use requestTls and proxyTls.

  113. CVE-2026-104827 7.5 Élevée

    reactphp http

    react/http is an event-driven, streaming HTTP client and server implementation for ReactPHP. From 1.0.0 until 1.11.1, React\Http\Browser uses ClientRequestStream::handleData() to append response-header bytes to an internal buffer without enforcing a size limit. A remote server can send a valid status line and partial header, withhold the \r\n\r\n or \n\n header terminator, and continue streaming data so the buffer grows for the life of the connection. Applications that fetch attacker-controlled or attacker-influenced URLs, including URLs reached through redirects, can exhaust PHP memory_limit and crash the event-loop or worker process. Confidentiality and integrity are unaffected. This issue is fixed in version 1.11.1.

  114. CVE-2026-55875 4.3 Moyenne

    HeliumEdu platform

    Helium is a free online student planner for classes, homework, grades, and notes. Prior to 2.2.60, any authenticated user can submit an arbitrary URL to POST /feed/externalcalendars/, where helium/feed/services/icalexternalcalendarservice.py uses validate_url to check only URL syntax before passing the value to urllib.request.urlopen, and fetch_ical_conditional periodically refetches the stored URL. The server can therefore issue requests to attacker-selected external or internal hosts, while distinct responses, including an uncaught http.client.RemoteDisconnected condition for non-HTTP services, reveal internal host and port reachability. The supported impact is forced outbound requests and internal service reconnaissance; the advisory does not demonstrate secret retrieval because the relevant ECS credentials path is task-specific and unpredictable. This issue is fixed in version 2.2.60.

  115. CVE-2026-55414 5.3 Moyenne

    nl-portal-backend-libraries

    NL Portal Backend Libraries provide backend components for Dutch government portals that interact with residents, customers, suppliers, and partner organizations. From 1.1.0 until 3.0.4, and in 3.1.0, the public getFormDefinitionByObjectenApiUrl and deprecated getFormDefinitionById GraphQL resolvers are reachable without CommonGroundAuthentication and forward a caller-influenced URL through ObjectenApiService.getObjectByUrl to ObjectsApiClient.getObjectByUrl, which attaches the privileged Objecten-API token. The host-equality check restricts requests to the configured Objecten-API host but does not constrain the scheme, port, or path, allowing an unauthenticated caller to induce a token-bearing request to a chosen endpoint on that host. Standard deployments are not exposed to cross-host SSRF, token capture requires attacker-controlled routing on the configured host, and typed deserialization limits returned data to objects shaped like form definitions. The 3.0 release line is fixed in version 3.0.4. The advisory does not identify a fixed 3.1 release.

  116. CVE-2026-103350 5.3 Moyenne

    Magepeople inc. Booking and Rental Manager

    Improper Validation of Specified Quantity in Input vulnerability in Magepeople inc. Booking and Rental Manager booking-and-rental-manager-for-woocommerce allows Removing Important Client Functionality.This issue affects Booking and Rental Manager: from n/a through 2.8.0.

  117. CVE-2026-102740 5.8 Moyenne

    AsyncHttpClient async-http-client

    The AsyncHttpClient library allows Java applications to execute HTTP requests and asynchronously process HTTP responses. From 2.0.0 until 2.16.1 and 3.0.12, PropertiesBasedResumableProcessor stores resumable download state at the predictable shared path $java.io.tmpdir/ahc/ResumableAsyncHandler.properties with default permissions and link-following file operations. A local user who can read the shared store can see download URLs, including embedded user information or query tokens. A local user who can write within or pre-create the affected shared store directory can plant a symbolic link that the client follows when saving state, truncating and overwriting a target file writable by the client process. The filesystem path is used when ResumableAsyncHandler decorates another AsyncHandler, while the no-argument constructors use a no-op processor and are unaffected. This issue is fixed in versions 2.16.1 and 3.0.12.

  118. CVE-2026-57568 5.7 Moyenne

    LycheeOrg Lychee

    Lychee is a free, open-source photo-management tool. Prior to 7.6.0, Lychee permits an authenticated user with album-edit permissions to store a javascript URI in an album or photo Markdown description because config/markdown.php enables allow_unsafe_links. app/Http/Resources/Models/Utils/PreformattedPhotoData.php and app/Http/Resources/Models/Utils/PreFormattedAlbumData.php convert the description to HTML, and the gallery components PhotoDetails.vue, Overlay.vue, PhotoThumb.vue, AlbumHero.vue, and AlbumCard.vue render that HTML with v-html. A visitor, including an unauthenticated visitor to a public album, who clicks the rendered link executes attacker-controlled JavaScript in the Lychee origin, which can expose session data or enable authenticated requests in the victim's context. This gallery-view flaw is distinct from CVE-2026-33738, which addressed feed rendering. This issue is fixed in version 7.6.0.

  119. CVE-2026-100353 7.5 Élevée

    getkin kin-openapi

    kin-openapi is a Go project for handling OpenAPI files. From 0.136.0 until 0.149.0, three openapi3filter decoder paths that were missed by the fix for GHSA-p6wj-qrr4-pgh5 lack a nil-items guard for legal OpenAPI 3.1 array schemas. An unauthenticated remote client can trigger a deterministic nil-pointer panic through an array query parameter, a deepObject query parameter containing an array property, or an application/x-www-form-urlencoded request body whose array property appears inside an allOf, anyOf, or oneOf subschema. The query and composed-body paths reach urlValuesDecoder.parseArray, while the deepObject path reaches buildResObj, and all pass a nil items schema into code that dereferences it. This issue is fixed in version 0.149.0.

  120. CVE-2026-97394 7.5 Élevée

    getkin kin-openapi

    kin-openapi is a Go project for handling OpenAPI files. From 0.136.0 until 0.143.0, openapi3filter accepts OpenAPI 3.1 array schemas without an items keyword and then dereferences the nil Schema.Items value while decoding application/x-www-form-urlencoded or multipart/form-data request bodies, response headers, and other array parameter surfaces. A remote unauthenticated client can reach an affected operation with a nonempty request or matching response and deterministically trigger a nil-pointer panic during ValidateRequest or ValidateResponse. OpenAPI 3.0 documents are not affected because document validation rejects arrays without items, and JSON request bodies use a different decoder path. This issue is fixed in version 0.143.0.

  121. CVE-2026-58393 5.9 Moyenne

    opnsense core

    OPNsense is a FreeBSD based firewall and routing platform. Prior to version 26.1.11, the OpenVPN Client Specific Override common_name field accepts path traversal sequences and openvpn_csc_conf_write() concatenates the value into /var/etc/openvpn-csc/<vpnid>/<common_name> without normalization or a basename check. The affected event handlers are client_connect.php and user_pass_verify.php. An authenticated user with Client Specific Override privileges can store the traversal value, and a matching OpenVPN client-connect or authentication event causes a generated CSO file to be written outside the intended directory. The file content is generated OpenVPN configuration rather than arbitrary bytes, and practical exploitation requires the attacker to cause a matching common name or username to be processed. This issue is fixed in version 26.1.11.

  122. CVE-2026-58390 8.0 Élevée

    opnsense core

    OPNsense is a FreeBSD based firewall and routing platform. Prior to version 26.1.11, the OpenVPN connection-status dashboard widget in src/opnsense/www/js/widgets/OpenVPNClients.js and the connection-status view place a client common_name into data-common-name attributes without quote-safe encoding. Exploitation requires an OpenVPN server configured with username-as-common-name and RADIUS or LDAP authentication, plus the ability to provision or self-register a valid backend username containing a double quote; deployments where the same administrator controls both the identity backend and firewall do not cross a separate trust boundary. An authenticated client using that username can close the attribute and add an event handler, while the local user database is not affected because its username validator rejects quotes. Script executes in an administrator's session when the administrator opens the dashboard or connection-status page and the injected event is triggered; the demonstrated payload used pointer movement over the client row rather than unconditional page-load execution. The advisory demonstrates that the script can read config.xml, invoke privileged APIs, reset the root password, and obtain root command execution. This issue is fixed in version 26.1.11.

  123. CVE-2026-111091 7.8 Élevée

    JetBrains Junie CLI

    In JetBrains Junie CLI before 26.8.3 code execution was possible via an untrusted project's MCP server configuration

  124. CVE-2026-110958 10.0 Critique

    Wavlink WN530HG4B

    A vulnerability was found in Wavlink WN530HG4B M30HG4B_V260724. The impacted element is an unknown function of the file api.cgi. The manipulation results in heap-based buffer overflow. It is possible to launch the attack remotely. The exploit has been made public and could be used. Upgrading to version 2026-09-24-841a51c6d5_tclinux.bin is sufficient to resolve this issue. The affected component should be upgraded. The vendor confirms: "We confirm the finding, and it has already been addressed. In the current firmware the api=info&client=app handler no longer builds a shell command: the login script is now invoked through an argument vector (execvp) instead of system(), and the client address is length-limited and null-checked."

  125. CVE-2026-100352 7.5 Élevée

    getkin kin-openapi

    kin-openapi is a Go project for handling OpenAPI files. Prior to 0.148.0, the OpenAPI 3.1 validation path passes YAML NaN and infinity numeric values to santhosh-tekuri/jsonschema v6.0.2 without applying kin-openapi's existing finite-number guards. A remote unauthenticated client can send an application/yaml or application/x-yaml request body containing .nan, .inf, or -.inf to an operation with a numeric constraint, causing the dependency to create and dereference a nil big.Rat during ValidateRequest. Standard net/http servers recover the panic per connection, while other hosts or application goroutines may terminate the process. This issue is fixed in version 0.148.0.

  126. CVE-2026-100350 5.3 Moyenne

    getkin kin-openapi

    kin-openapi is a Go project for handling OpenAPI files. From 0.145.0 until 0.146.0, openapi3filter repeatedly calls regexp.MustCompile for the same parameter-name regular expression inside the per-query-key loop in urlValuesDecoder.DecodeObject used to decode a deepObject query parameter. An unauthenticated remote client can send an affected endpoint a request containing many unrelated query keys, causing ValidateRequest to perform work and allocations proportional to the attacker-controlled key count for every deepObject parameter. The expression itself is fixed and this is not regular expression denial of service; the resource consumption comes from loop-invariant recompilation. This issue is fixed in version 0.146.0.

  127. CVE-2026-100347 5.3 Moyenne

    getkin kin-openapi

    kin-openapi is a Go project for handling OpenAPI files. Prior to 0.145.0, openapi3filter.ValidationHandler uses the routers/legacy router, which passes an unsupported or non-canonical HTTP method for an existing static path to openapi3.PathItem.GetOperation. The default branch of GetOperation panics before authentication runs, so an unauthenticated remote client can repeatedly send methods such as PROPFIND or lowercase get to abort connections and generate stack-trace log growth. The specific panic requires an exact static path; a concrete path matched only by a templated route does not reach the same sink. This issue is fixed in version 0.145.0.

  128. CVE-2026-95180 5.5 Moyenne

    An issue in Justson agentweb-core v5.1.6-androidx and earlier allows a local attacker to obtain sensitive information and launch non-exported host application components via incorrect access control in the handling of intent:// URLS in DefaultWebClient

  129. CVE-2026-95179 7.5 Élevée

    An issue in lottie-android 6.7.1 and earlier (and master) allows a remote attacker to cause a denial of service via a crafted Lottie animation JSON containing a cyclic precomposition reference graph that triggers uncontrolled recursion in BaseLayer.forModel and CompositionLayer.

  130. CVE-2026-59994 4.3 Moyenne

    espocrm

    EspoCRM is an open source customer relationship management application. Prior to 9.3.9, an unauthenticated remote attacker can send an email to a monitored group or IMAP address, including an email-to-Case address, and place arbitrary HTML in its plain-text body. EspoCRM stores that body and renders it without escaping in Email Received stream notification emails sent to users following the related record, allowing HTML injection and, in mail clients that execute it, JavaScript execution. Earlier sanitization for Post and Mention notifications did not cover this notification path. This issue is fixed in version 9.3.9.

  131. CVE-2026-55184 3.5 Faible

    espocrm

    EspoCRM is an open source customer relationship management application. Prior to 9.3.9, a user with Email Template or PDF Template edit permission can store crafted HTML that bypasses moderateSanitizeHtml(). When another user opens the template in edit mode and clientCspDisabled is true, the stored payload executes and can hijack the user's session. This issue is fixed in version 9.3.9.

  132. CVE-2026-110957 10.0 Critique

    Wavlink WL-WN530BE2-A

    A vulnerability has been found in Wavlink WL-WN530BE2-A up to M30BE2_V250604. The affected element is the function sub_12580 of the file api.cgi of the component Login Processing. The manipulation of the argument REMOTE_ADDR leads to os command injection. It is possible to initiate the attack remotely. The exploit has been disclosed to the public and may be used. Upgrading to version 2026-09-24-841a51c6d5_tclinux.bin is sufficient to fix this issue. You should upgrade the affected component. The vendor confirms: "We confirm the finding, and it has already been addressed. In the current firmware the api=info&client=app handler no longer builds a shell command: the login script is now invoked through an argument vector (execvp) instead of system(), and the client address is length-limited and null-checked."

  133. CVE-2026-93649 7.5 Élevée

    Eclipse Foundation Adoptium Temurin

    In the Eclipse Adoptium website (adoptium.net), in every deployment built from commit 0b24ea1 (2025-06-18) up to the fix in commit cbb53cf (2026-09-24), the post-download page /[locale]/download takes a download URL from the "link" query parameter and automatically navigates to it with a meta http-equiv="refresh". The URL is checked against a list of trusted prefixes using String.startsWith() on the raw string, while the browser applies RFC 3986 dot-segment normalisation before navigating, so a value such as https://github.com/adoptium/temurin/../../attacker/malware/releases/download/v1/jdk.exe passes the check but resolves to an attacker-controlled repository. A victim who follows a crafted adoptium.net link is shown the genuine "Thank you for downloading Eclipse Temurin" page while their browser downloads a binary of the attacker's choosing

  134. CVE-2026-110724 7.5 Élevée

    Eclipse Foundation Eclipse Connected Services Platform

    In `api-gateway` versions till `1.7.3`, The default rate-limit key resolver, `ClientIpKeyResolver`, derives the per-client bucket key from the leftmost element of the `X-Forwarded-For` request header. That element is supplied by the originating HTTP client and is not validated against any trusted-proxy list. A single attacker can therefore obtain an unlimited number of independent rate-limit buckets — one per unique header value — and completely bypass the gateway's request-rate limiting on every route that uses the shipped default configuration.

  135. CVE-2026-110723 7.5 Élevée

    Eclipse Foundation Eclipse Connected Services Platform

    In `api-gateway` versions till `1.7.3`, The default rate-limit key resolver, `ClientIpKeyResolver`, derives the per-client bucket key from the leftmost element of the `X-Forwarded-For` request header. That element is supplied by the originating HTTP client and is not validated against any trusted-proxy list. A single attacker can therefore obtain an unlimited number of independent rate-limit buckets — one per unique header value — and completely bypass the gateway's request-rate limiting on every route that uses the shipped default configuration.

  136. CVE-2026-104025 8.4 Élevée

    PeaZip

    Path Traversal (CWE-22) in the execute_obj function of the archive object viewer (peach.pas) in PeaZip 11.2.0 and earlier on Windows allows an attacker who convinces a victim to preview an entry of a crafted archive via the "Preview with... -> Associated application" action (Shift+Ctrl+O) to execute an arbitrary program already present on the victim's system, with no warning dialog. The function validates only outname (PeaZip's own temporary directory) with checkfiledirname, while the string actually executed is the concatenation outname+tempstring; tempstring is taken verbatim from the archive entry name (StringGridList.Cells[12,...]) and is never validated. checkfiledirname (list_utils.pas) rejects control and reserved characters in the file name but performs no ".." path-containment check, so an entry named ../../../../Windows/System32/calc.exe escapes the extraction directory and is passed to cp_open -> ShellExecuteW; the extracting backend (7-Zip) strips the traversal while the opener does not. In addition, the SECEXTCONST dangerous-extension confirmation prompt is implemented only in the double-click handler (grid_obj_open) and not on this preview route, so an executable inside an archive reaches ShellExecuteW unannounced, and where the prompt does fire it names the entry's display name rather than the resolved path that will run. Arbitrary code execution was demonstrated with a proof of concept against official PeaZip builds, confirmed on two archive formats (.zip and .tar) to show the defect is in PeaZip's own path handling and not in a single format parser.

  137. CVE-2026-110726 6.5 Moyenne

    Eclipse Foundation Eclipse Connected Services Platform

    In `api-gateway` versions till `1.7.3`, The default rate-limit key resolver, `ClientIpKeyResolver`, derives the per-client bucket key from the leftmost element of the `X-Forwarded-For` request header. That element is supplied by the originating HTTP client and is not validated against any trusted-proxy list. A single attacker can therefore obtain an unlimited number of independent rate-limit buckets — one per unique header value — and completely bypass the gateway's request-rate limiting on every route that uses the shipped default configuration.

  138. CVE-2026-19864 5.6 Moyenne

    Eclipse Jetty's `SessionAuthentication` retains authentication credentials in a serializable object. When the credentials contain a plaintext password and the authenticated session is persisted to a session data store, the password may consequently be written to persistent storage in plaintext. An attacker with sufficient privileges to read the persisted session data may obtain the plaintext passwords of authenticated users.

  139. CVE-2026-19863 9.3 Critique

    An unauthenticated client may access certain protected resources when using EthereumAuthenticator if the request path contains the configured login path. This is caused by the authenticator using an unanchored substring match when identifying the login page. If the login path appears later within a protected request path, the request may be treated as the login page and its authentication or role requirement replaced with ALLOWED. Exploitation requires the protected resource path to contain the configured login path as a later segment. Explicitly FORBIDDEN routes are not affected.

  140. CVE-2026-18662 6.9 Moyenne

    In Eclipse Jetty, across various versions of `jetty-security`, and `jetty-eeN-security`, two or more security constraints with the same http-method-omission set, and no url-patterns, are not combined properly per Jakarta Servlet Spec. Resulting in later http-method-omission sets overriding previous constraints with the same http-method-omission set.

  141. CVE-2026-18660 8.8 Élevée

    In Eclipse Jetty, across various versions of `jetty-server`, and `jetty-proxy`, the host whitelist/blacklist is case-sensitive, and would allow hosts through that only differ by case.

  142. CVE-2026-18659 6.9 Moyenne

    In Eclipse Jetty, across various versions of `jetty-security`, and `jetty-eeN-security`, and `jetty-eeN-servlet`, the Jakarta Spec constraints defined as `http-method-omission` can be bypassed by using substrings of the defined HTTP Methods.

  143. CVE-2026-77711 8.2 Élevée

    Eclipse Foundation Eclipse Jetty

    When Jetty's `HttpClient` make a request to a server through a forward proxy, the server may reply with a 407 (Proxy Authentication Required). The client mistakenly interprets the 407 as coming from the proxy, so it re-sends the request with the `Proxy-Authentication` header, which arrives to the server, that now has access to the proxy credentials.

  144. CVE-2026-76184 8.2 Élevée

    Eclipse Foundation Eclipse Jetty

    In Eclipse Jetty, HttpClient may receive a redirect of the form: ``` /secure/../evil ``` If an attacker is able to control the redirects, then HttpClient matches authentication credentials against `/secure` rather than normalizing the path to `/evil`, and will therefore follow the redirect sending the `Authorization` header for `/secure` to `/evil`. This also mean that the attacker must have an evil application deployed on the same domain as the target.

  145. CVE-2026-76182 6.9 Moyenne

    Eclipse Foundation Eclipse Jetty

    In Eclipse Jetty, `TryPathsHandler` and `TryFilesFilter` can be configured to forward requests to template URIs such as `index?p=$path`, where `$path` is the request path in context. However, URI paths may contain characters that are reserved in URI queries so that a path of `/s&a=1` when applied to the template becomes `index?p=/s&a=1`, which is interpreted as 2 query parameters: `p=/s` and `a=1`, rather than just one query parameter `p` whose value is `/s&a=1`.

  146. CVE-2026-18661 8.8 Élevée

    In Eclipse Jetty, across various versions of `jetty-annotations`, and `jetty-eeN-annotations`, the Jakarta Spec authoritative url-pattern rules for mixed `@ServletSecurity` annotation, and `web.xml` defined `security-constraint` on the same `url-pattern` were not being applied if the `web.xml` had more than 1 `url-pattern` for a servlet, and had a `security-constraint` applying to only some of those `url-pattern` definitions.

  147. CVE-2026-109855 7.5 Élevée

    undici · Nodejs undici

    undici is an HTTP client library for Node.js. In versions 8.11.0 and 8.11.1, its experimental WebSocket client, which is enabled by default when HTTP/2 is allowed, mishandles a failed WebSocket handshake over HTTP/2: the request is already marked completed, so the abort becomes a no-op and the underlying HTTP/2 stream is left open with no error listener. When a malicious server then destroys the connection, the orphaned stream emits an unhandled error and the Node.js process exits. A remote, unauthenticated server can therefore crash any client that opens a WebSocket to it over HTTP/2, resulting in a denial of service. The issue is fixed in undici 8.11.2, and users should upgrade to 8.11.2 or later. As a workaround, set allowH2 to false to disable HTTP/2 for WebSocket connections, or avoid opening WebSocket connections to untrusted servers.

  148. CVE-2026-103224 5.9 Moyenne

    Eclipse Foundation Eclipse Vert.x · Eclipse vert.x

    In Eclipse Vert.x versions [4.0.0, 4.5.34] or [5.0.0,5.0.12] or [5.1.0,5.1.9] or [5.2.0, 5.2.0], the MySQL message decoder aggregates inbound multi-packet payloads into a composite direct-memory buffer without any cumulative size limit. A malicious, compromised, or man-in-the-middle MySQL server can send a continuous stream of packets with the maximum payload length (0xFFFFFF), preventing the end-of-message condition from being reached, causing unbounded direct-memory allocation on the client and leading to an out-of-memory crash. This occurs before authentication completes. The impacted component is io.vertx:vertx-mysql-client.

  149. CVE-2026-103223 5.9 Moyenne

    Eclipse Foundation Eclipse Vert.x · Eclipse vert.x

    In Eclipse Vert.x versions [4.0.0, 4.5.34] or [5.0.0,5.0.12] or [5.1.0,5.1.9] or [5.2.0, 5.2.0], the PostgreSQL message decoder reads the 4-byte message length field from the PostgreSQL wire protocol without any maximum validation. A malicious, compromised, or man-in-the-middle PostgreSQL server can send a message header with the length field set to an extremely large value (up to Integer.MAX_VALUE), causing the decoder to accumulate subsequent TCP data into a composite buffer without bound, leading to unbounded memory allocation on the client and an out-of-memory crash. This occurs before authentication completes. The impacted component is io.vertx:vertx-pg-client.

  150. CVE-2026-102477 5.9 Moyenne

    Eclipse Foundation Eclipse Vert.x · Eclipse vert.x

    In Eclipse Vert.x versions [4.0.0, 4.5.34] or [5.0.0,5.0.12] or [5.1.0,5.1.9] or [5.2.0, 5.2.0], a gRPC endpoint does have a configurable per message maximum size in bytes, however the maximum size applies to the received bytes. When applied to a compressed message, this only applies to the compressed size and not the actual size of the uncompressed content. This can be used to force a gRPC endpoint to allocate more bytes than the value defined by the configuration, making the endpoint vulnerable to compression bomb attacks. The impacted component is io.vertx:vertx-gprc-server.