The npm package service-roles-registry version 99.0.2 contains malicious code that executes a preinstall script. This script collects host identifiers such as hostname, user info, current working directory, and package metadata, then sends this data to a hardcoded external endpoint without user consent. The package appears to be a dependency confusion lure designed to override an internal package of the same name.
Tamery versions from 0.26.1 up to but not including 0.99.0 contain an authorization bypass vulnerability in the cloud sync API's chatsMessages.update handler. This flaw allows authenticated Pro subscribers to manipulate the client-supplied chatId parameter, enabling them to move messages with arbitrary content into another user's AI chat history. The vulnerability is fixed in version 0.99.0.
Tamery versions from 0.32.0 up to but not including 0.99.0 contain an authorization bypass vulnerability in the cloud sync API's connectionsResources.create handler. This flaw allows authenticated users to attach resources to other users' connections by including a victim's connectionId in batch requests. The issue is fixed in version 0.99.0.
The Rust compiler executable (rustc) version 1.99.0 for Windows has been flagged as malicious by multiple antivirus vendors on VirusTotal. This detection appears specific to version 1.99.0, as version 1.99.1 is not flagged. The issue was reproducible with a fresh installation of Rust 1.99.0 via rustup. The SHA256 hash of the flagged binary is FC40F1C8B65095A22A48C111AADC279C26A049B7E4161E8B99FFE3EB23873E86. No further technical details or confirmed exploits are available.
In the Linux kernel, the following vulnerability has been resolved:
ntfs3: free the sbi after a grace period
If the "nocase" mount option ntfs3 uses custom hash and dentry compare
functions. Both access ntfs3 superblock private information in rcu
pathwalk. ntfs_put_super() and ntfs3_kill_sb() drop the info without any
rcu delay. A lookup that's still inside the filesystem when the last
reference to the mount is dropped causes a UAF. With a pathwalk held up
in ntfs_d_hash() while the last reference to the mount goes away:
BUG: KASAN: slab-use-after-free in ntfs_nls_to_utf16+0x752/0x940
Read of size 8 at addr ff1100010453c850 by task walker/99
Call Trace:
<TASK>
ntfs_nls_to_utf16+0x752/0x940
ntfs_d_hash+0x22b/0x500
link_path_walk+0x11ad/0x1d30
path_lookupat+0x63/0xb00
filename_lookup+0x1f1/0x5b0
vfs_statx+0xd7/0x3b0
[...]
Allocated by task 95:
__kmalloc_cache_noprof+0x249/0x560
ntfs_init_fs_context+0x2db/0x5a0
alloc_fs_context+0x52e/0xc70
[...]
Freed by task 98:
kfree+0x1db/0x5b0
deactivate_locked_super+0x9c/0xf0
cleanup_mnt+0x9f/0x130
task_work_run+0x131/0x250
The bug is currently masked by the grace period in namespace_unlock()
for umount(2). Keep the necessary information until the filesystem
private info is freed and free it with an rcu delay.
ntfs3_free_sbi() now runs from an rcu callback and drops the upcase
table via ntfs_put_shared(). So switch s_shared_lock over to
spin_lock_bh(). Otherwise the callback can deadlock against a task
holding the lock on the same cpu.
In the Linux kernel, the following vulnerability has been resolved:
net: psp: do not inherit the Rx association on clone
sk->psp_assoc sits past sk_dontcopy_end, so sock_copy() copies it into
every socket accepted from a listener without taking a reference, while
inet_sock_destruct() puts for every inet socket. psp_twsk_init() does
refcount_inc() for the timewait socket, so a child closing through
TIME_WAIT cancels its own put and leaves the association with one
reference and N timewait sockets holding the same pointer. Closing the
listener frees it, and the timewait timers then put freed memory.
Rejecting the association on a listening socket is not sufficient: a socket
can acquire one while established and then be turned back into a listener,
because tcp_disconnect() leaves sk->psp_assoc in place.
BUG: KASAN: slab-use-after-free in psp_twsk_assoc_free+0x6f/0xf0
Write of size 4 at addr ffff888110f9255c by task swapper/7/0
psp_twsk_assoc_free+0x6f/0xf0
inet_twsk_put+0xda/0x1b0
call_timer_fn+0x53/0x2e0
__run_timers+0x764/0xa80
Freed by task 99:
kfree+0x1a7/0x500
process_one_work+0x7ec/0x1100
An association carries a per-connection SPI and key, so a child must not
inherit the parent's. Clear it on clone.
The npm package '@capsense/core-ui' version 99.99.100 has been identified as malicious by the OpenSSF Package Analysis project. The package is flagged because it communicates with a domain associated with malicious activity. There is no CVSS score available, and no known exploits in the wild have been reported. No patch or remediation information is provided.
CVE-2026-114726 is a mail header injection vulnerability in Roundcube Webmail affecting versions before 1.6.20 and 1.7.5. The flaw exists in the Bounce (resend) action where control characters in the From display-name are not properly stripped, allowing injection of arbitrary headers. This issue was fixed in versions 1.6.20 and 1.7.5.
CVE-2026-114725 is a mail header injection vulnerability in Roundcube Webmail versions before 1.6.20 and 1.7.x before 1.7.5. The function rcmail_sendmail only sanitized newline characters in the Subject and Organization headers, leaving other headers such as References, In-Reply-To, Message-ID, and From vulnerable to injection. This allows a remote authenticated attacker to inject arbitrary headers into outgoing emails. The issue was fixed by normalizing whitespace in all string header values in versions 1.6.20 and 1.7.5.
CVE-2026-114717 is a mail header injection vulnerability in Roundcube Webmail versions before 1.6.20 and 1.7.5. The flaw occurs when forwarding a message as an attachment, where the subject line is used directly to name the attachment. If the subject contains CRLF sequences, it can inject arbitrary MIME headers into the outgoing message. This issue was fixed by collapsing whitespace in the generated attachment name in versions 1.6.20 and 1.7.5.
CVE-2026-114716 is a mail header injection vulnerability in Roundcube Webmail versions before 1.6.20 and 1.7.x before 1.7.5. The flaw arises from improper validation of the media type in inline image data URIs, allowing a remote authenticated attacker to inject arbitrary MIME part headers into outgoing messages. This issue was fixed by validating the media type with rcube_mime::is_mimetype_valid().
Parse Server versions before 8.6.99 and versions 9.0.0 up to but not including 9.10.3-alpha.2 contain an incorrect authorization vulnerability. The server evaluates checks against the target account during user updates before verifying the caller's authorization, leaking information about usernames, emails, and, under certain password policy settings, current or previous passwords. This allows an unauthorized attacker to confirm guessed credentials and bypass account lockout mechanisms.
CVE-2026-49206 is an integer overflow or wraparound vulnerability in the rapidcsv C++ CSV parser library versions prior to 8.99. The issue arises from converting size_t indexes to int before adding them to std::vector iterators on 64-bit systems. When handling CSV matrices with more than 2^31 elements, an index exceeding INT_MAX can wrap to a negative offset, causing undefined behavior, heap memory corruption, or a crash. Exploitation requires extremely large in-memory CSV documents, making typical use cases unlikely to be affected. The vulnerability is fixed in version 8.99.
CVE-2026-50218 is a high-severity vulnerability in Akamai Guardicore Platform Agent (GPA) versions before 7.4. The GPA client improperly validates JWT tokens used for seeding, allowing an attacker to inject arbitrary parameters via a spoofed JWT token. This can lead to unauthorized reconfiguration and full compromise of the application's functionality. The attack requires convincing a user to interact with a malicious link or knowledge of the IDP URL used by an enrolled client. A patch is available to address this issue.
The npm package react-paginate-v5@99.0.0 is a malicious package designed as a dependency confusion probe. It masquerades under a name similar to the legitimate react-paginate package but uses an implausibly high version number to override internal packages. Upon installation, it executes a preinstall script that performs a DNS lookup to an attacker-controlled domain, confirming the package's installation and revealing resolver metadata. The package has no legitimate functionality and can potentially execute arbitrary shell commands during installation.
The npm package 'react-dropzone-v4' version 99.0.0 is a malicious package that mimics the popular 'react-dropzone' library. It contains a preinstall script that executes automatically during installation, which performs DNS and HTTP requests to a third-party domain, transmitting the installer's hostname without consent. The package provides no legitimate functionality and appears designed for dependency confusion attacks to capture host information.
CVE-2026-108593 is a high-severity code injection vulnerability in decolua 9router versions 0.4.1 through 0.5.99. It affects the POST /api/cli-tools/hermes-settings endpoint, allowing authenticated dashboard users to inject arbitrary keys into the Hermes Agent config.yaml file. By submitting a specially crafted baseUrl containing double quotes and newlines, attackers can add a hooks_auto_accept setting and a hooks.post_llm_call shell command. This command is executed by the Hermes Agent without user approval after an LLM call, potentially enabling unauthorized code execution.
The npm package '@library-wide/library-shell' version 99.9.1 has been identified as malicious due to its communication with a domain linked to malicious activity. This behavior indicates the package may be used to facilitate harmful operations. No CVSS score is available for this threat.
Magic: The Gathering Arena (Windows/Steam client) versions 2026.59.30.12801.127931.6 and certain later 2026.60.x builds have a vulnerability where a server-supplied URL from the home-screen carousel is passed directly to the Windows shell without validating the URI scheme or domain. This can allow an attacker controlling carousel content to invoke arbitrary registered URI-scheme handlers on client hosts without user interaction. The vulnerability is classified as CWE-99 (Improper Control of Resource Identifiers) and has a low CVSS score of 3.4.
CVE-2026-106583 is a low-severity vulnerability in OpenSSH before version 10.6 where the presence of a $ or \ character in a command-line username can lead to resource injection. This improper control of resource identifiers could allow injection attacks via crafted usernames. The vulnerability has a CVSS score of 2.5, indicating low impact, with limited integrity impact and no confidentiality or availability impact.