Out-of-Bounds Write in Android s_gen_tile_param Leading to Remote Denial of Service

In s_gen_tile_param of multiple files, there is a possible way to trigger an out of bounds write in tile parameter arrays due to an incorrect bounds check. This could lead to remote denial of service with no additional execution privileges needed. User interaction is not needed for exploitation.

Android Vulkan API Use After Free in Image View Creation Allows Remote Code Execution

In IMG_vkCreateImageView of image_view.c, there is a possible way to trigger a use after free due to memory corruption. This could lead to remote code execution with no additional execution privileges needed. User interaction is not needed for exploitation.

Android bytes_are_equal Side-Channel Leakage Exposes Auth Keys

In bytes_are_equal of lib.rs, there is a possible way to recover authentication keys due to side channel information disclosure. This could lead to local information disclosure with no additional execution privileges needed. User interaction is not needed for exploitation.

Out‑of‑Bounds Write in ExecuteGraphCommand Enables Local Privilege Escalation

In ExecuteGraphCommand of multiple files, there is a possible way to trigger an out of bounds write via an oversized scratch memory declaration due to an incorrect bounds check. This could lead to local escalation of privilege with no additional execution privileges needed. User interaction is not needed for exploitation.

Laravel: Task Output Injection - Concurrency Process

Laravel is a web application framework. Prior to 12.71.0 and 13.40.0, Laravel's Concurrency::run process driver can treat a forged internal result message in child standard output as framework output. An application route must invoke Concurrency::run, and a task must write attacker-controlled bytes at the start of the child's standard output that form a complete JSON document followed by gzip magic bytes. Under those application-specific conditions, an attacker can forge the internal result message and achieve remote code execution. The affected behavior is not exposed by default and requires application code that routes untrusted input into task output. This issue is fixed in versions 12.71.0 and 13.40.0.

BACnet Stack: Pre-auth OOB read in xy_color_decode (BACnetXYColor) via WriteGroup/WriteProperty

BACnet Stack is a BACnet open source protocol stack C library for embedded systems. Prior to versions 1.4.5 and 1.5.1, a single unauthenticated BACnet/IP UDP datagram (port 47808) triggers an out-of-bounds read in xy_color_decode() (src/bacnet/lighting.c), reachable pre-authentication through two default-registered service handlers: the UnconfirmedRequest WriteGroup service (0x0A) and Confirmed WriteProperty on a BACnetXYColor-typed property. This issue has been patched in versions 1.4.5, 1.5.1, and 1.6.0.

OpenTelemetry-Go: Eager W3C baggage-property materialization causes bounded memory-allocation amplification

OpenTelemetry-Go is the Go implementation of OpenTelemetry. From version 1.0.0-RC1 until version 1.47.0, baggage.Parse and propagation.Baggage.Extract eagerly materialize every W3C baggage property through parseMember, grow property slices, serialize members, and copy properties into internal/baggage.Item. A valid densely populated baggage header can therefore cause allocation traffic and retained memory disproportionate to its bounded input size. Remote impact requires baggage extraction to be enabled, attacker-controlled input to reach it, and extraction to run before authentication or on a public route; the evidence establishes increased allocation and garbage-collection pressure but not reliable process termination. This issue is fixed in version 1.47.0.

russh kex_done: shared-secret callback runs before server host-key verification

russh is a Rust SSH client and server library that includes an SSH agent protocol implementation. From version 0.55.0 until version 0.65.1, the initial key exchange invokes client::Handler::kex_done() with the raw shared secret before check_server_key() accepts the server host key or certificate. A malicious SSH server can therefore trigger application-defined secondary-channel initialization or early data handling even when the server identity is subsequently rejected. Ordinary clients that keep the callback's no-op default or independently defer callback-derived actions until trust is confirmed are not affected. This issue is fixed in version 0.65.1.

bgpd: out-of-bounds heap read and crash on NOTIFICATION Cease/Hard Reset with one data byte

In FRRouting (FRR) 8.3 through 10.5.1, bgp_notify_decapsulate_hard_reset() in bgpd/bgp_packet.c does not validate the length of the data field of a received BGP NOTIFICATION Cease/Hard Reset message (RFC 8538). When the data field is exactly one byte long, the function reads one byte past the end of a heap buffer and the 16-bit length of the encapsulated data underflows to 65535, causing a 65535-byte out-of-bounds heap read followed by an assertion failure. A remote BGP peer that has established a session and negotiated the Graceful Restart capability with the Notification (N) bit can send such a message to crash the bgpd daemon, terminating all BGP sessions (denial of service). This is fixed in 10.0.5, 10.1.5, 10.2.6, 10.3.4, 10.4.3, 10.5.2 and 10.6.0.

russh try_sign: asynchronous confirmation can sign after agent lock

russh is a Rust SSH client and server library that includes an SSH agent protocol implementation. Prior to version 0.65.1, try_sign() captures an Arc<PrivateKey> before awaiting Agent::confirm() and does not revalidate the shared lock state or key presence after the await. This is a distinct residual defect in the remediation for CVE-2026-116794. A peer, including a remote SSH server using agent forwarding, can begin a confirmation-constrained signature request while the agent is unlocked and receive the signature after another connection locks the agent if the stale confirmation returns true. New post-lock requests are rejected and private key material is not disclosed, but the pending operation bypasses the current lock state. This issue is fixed in version 0.65.1.

dnspython excessive memory and CPU use

dnspython is a DNS toolkit for Python. Prior to version 2.9.2, the from_text() parser for WKS records does not require the service number to be between 0 and 65535 before using it as a bitmap offset. An attacker-controlled zone file can supply a larger number, causing the service bitmap to grow to an extremely large size through inefficient repeated extension, consuming excessive memory and CPU and potentially terminating the process. Applications are affected only when they parse zone-file text whose record type and rdata are attacker controlled. This issue is fixed in version 2.9.2.

SvelteKit: Protocol-relative module import during server-side route resolution allows cross-origin script execution

SvelteKit is a framework for rapidly developing robust, performant web applications using Svelte. Prior to versions 2.70.4 and 3.1.1, apps configured with router.resolution: 'server' could treat a pathname beginning with // as a protocol-relative module specifier during client-side navigation. A socially engineered link could therefore make server-side route resolution import and execute JavaScript from another origin in the user's browser. This issue is fixed in 2.70.4 and 3.1.1.

SvelteKit: Unhandled URIError when `reroute` rewrites a malformed pathname crashes adapter-node

SvelteKit is a framework for rapidly developing robust, performant web applications using Svelte. Prior to versions 2.70.4 and 3.1.1, the reroute hook could rewrite an invalid percent-encoded pathname into a valid route and reach a comparison that decoded the original malformed pathname outside the guarded decoding operation. The resulting unhandled URIError could crash an adapter-node server and deny service to other requests. This issue is fixed in 2.70.4 and 3.1.1.

SvelteKit: Build-time server-only module import guard could fail, resulting in server secrets being statically replaced in client bundle

SvelteKit is a framework for rapidly developing robust, performant web applications using Svelte. From version 3.0.0 until 3.1.1, the build-time server-only module guard did not treat src/params.js as a client entry point even though parameter matchers can run in the browser. Server-only modules imported through src/params.js could consequently enter the client bundle, where statically replaced private environment values or other server secrets could be exposed. This issue is fixed in 3.1.1.

SvelteKit: Private `query.batch` functions could collide, producing unexpected parameter / function combinations

SvelteKit is a framework for rapidly developing robust, performant web applications using Svelte. From version 2.60.1 until 3.1.1, applications using the experimental remote functions flag could assign the same generated identifier to separate non-exported query.batch functions. When those private functions shared a request batch, arguments intended for one function could be passed to another function, causing unexpected cross-function processing and possible disclosure of function arguments. This issue is fixed in 3.1.1.

SvelteKit: DoS while parsing form data from a remote functions `form`

SvelteKit is a framework for rapidly developing robust, performant web applications using Svelte. From version 2.42.0 until 3.1.1, applications using the experimental remote functions flag and a remote form function could perform quadratic synchronous work while parsing a specially crafted request with many form entries. The parsing work could block the event loop for an extended period and deny service. This issue is fixed in 3.1.1.

Vault vulnerable to ACL deny bypass via case-variant auth method user and group names

Vault Community Edition and Vault Enterprise did not apply explicit ACL policy deny rules to letter case variants of user and group names in auth methods that treat these names as case-insensitive. This may allow an authenticated user with delegated permissions to modify a protected user or group, potentially leading to privilege escalation. This vulnerability (CVE-2026-128100) is fixed in Vault Community Edition 2.2.2 and Vault Enterprise 2.2.2, 1.21.14, and 1.19.25.

Abrt: abrt: abrt-server trusts client-supplied pid= allowing unprivileged read of other processes' /proc environ

A flaw was found in abrt. abrt-server trusts a client-supplied PID when creating problem dumps over abrt.socket and, running as root, copies that process's /proc data (including environ) into a dump directory owned by the connecting unprivileged user. This allows a local attacker to read environment variables and related metadata of other users' or root processes when abrtd is running.

Abrt: abrt: incomplete cve-2026-54228/54229 fix — completeness gate bypassed via attacker-supplied count element

A flaw was found in abrt. The completeness check introduced to mitigate CVE-2026-54228 and CVE-2026-54229 only verifies that a dump directory contains a "count" element, which an unprivileged user can create. This allows a local attacker to bypass the guard and again poison elements in, or seize ownership of, dump directories while privileged ABRT handlers are processing them. Related Problems2 mutating methods remain ungated by the same check.

Incorrect default permissions in the agent configuration folder in Devolutions PowerShell Universal 2026.3.4 and earlier allows a local unprivileged user to execute arbitrary code as SYSTEM via a crafted agent configuration file that points the agent to an attacker-controlled server.