An out-of-bounds heap read in Furnace's FamiTracker module importer (through 0.6.8.3) allows a crafted .ftm, .dnm or 0CC/E-FamiTracker file to drive DivEngine::loadFTM() past the noteMap array when a 2A03 instrument carries a DPCM note count above 120, crashing the tracker when the file is opened. The flaw is rated CVSS 6.8 (assessed as CVSS:3.1 AV:L/AC:L/PR:N/UI:R with C:N/I:N/A:H), reflecting that it is a local, file-based attack requiring the victim to open the malicious module; the documented outcome is a crash, not code execution or data disclosure. No confirmed active exploitation (not in CISA KEV, SSVC exploitation: none) and a low EPSS of 0.14% (3rd percentile) place it in the this_month remediation tier; a public technical writeup describing the import bug exists, but no verified working exploit code was identified at time of analysis.
Heap out-of-bounds read in libfvde through 20260923 can crash any application that opens a crafted FileVault 2 Core Storage volume via libfvde_volume_open(). The type 0x0024 encrypted-metadata parser uses an xml_plist_data_size value without checking it against the source metadata block, so a malicious volume image causes a read past the heap buffer and terminates the parsing process, with availability impact only and no confidentiality or integrity impact. The CVSS 4.0 vector is local (AV:L) with no privileges required (PR:N) and passive user interaction (UI:P); EPSS is 0.14% (3rd percentile) and no CISA KEV entry is provided, though a public PoC reference exists and is not verified as working exploit code.
An out-of-bounds read in gbsplay 0.0.103 and earlier lets a crafted Game Boy sound file crash the player: mapper_map() in mapper.c accepts a ROM bank index larger than the actual ROM, computing an underflowed bank length that bank_get() then reads past the end of the heap-allocated ROM buffer. The attack is local and requires the victim to open a malicious .gbr, .gbs or .gb file (CVSS 4.0 6.8; PR:N but UI:P), and the confirmed impact is a denial of service against the player process rather than code execution or disclosure of data back to the attacker. No public exploit code was identified at time of analysis, and EPSS is low at 0.14% (3rd percentile); an upstream fix commit exists, but a tagged release containing it is not independently confirmed. This is a low-to-moderate real-world priority despite the 6.8 CVSS 4.0 score.
libmsiecf, the libyal library for parsing Microsoft Internet Explorer cache index.dat files, contains a double free in libmsiecf_url_values_read_data() that corrupts heap memory when a recovered partial URL record carries an oversized unknown_offset, allowing a crafted index.dat file to crash the forensic tool parsing it. All versions through 20260525 are affected, and the impact is availability-only: there is no privilege escalation or data disclosure, the vector is local, and an operator must actually ingest the malicious file (CVSS 4.0 6.8; independently assessed as CVSS 3.1 AV:L/AC:L/PR:N/UI:R with A:H). Exposure is concentrated in DFIR/forensic pipelines that process untrusted evidentiary cache files; EPSS is 0.13% (2nd percentile), no public exploit code has been confirmed at time of analysis, and a public technical write-up (HackMD) describes the flaw. This is a genuine but low-urgency issue, not a high-severity emergency.
Authorization-code injection in the generic OpenID Connect backend of social-auth-core lets an attacker who holds a victim's unused OIDC authorization code redeem it inside their own browser and be authenticated as the victim. The nonce is stored in a shared OpenIdConnectAssociation table and is only checked for existence and expiry, not for binding to the browser session and OAuth state that started the login, so the callback succeeds even though the state belongs to the attacker; the flaw is exploitable only with the default DEFAULT_USE_PKCE = False. Affected are social-auth-core releases before 6.2.0 (reproduced on 6.1.1, 6.1.0 and 5.2.0, code identical on master) including every backend derived from OpenIdConnectAuth; CVSS 6.8 reflects that the attacker must first obtain a live victim code and lure the victim through a login (AC:H, UI:R). A public PoC reference is published with the GHSA advisory, no CISA KEV entry or confirmed active exploitation exists, and the vendor fixed the issue in 6.2.0.
boneIO Black Debian images ship the Mosquitto MQTT broker password file at /etc/mosquitto/passwd with world-readable permissions, so any account with local shell access can read the hashed credentials for every MQTT account configured on the device. This is a moderate, locally-scoped information-disclosure issue (CWE-732, incorrect permission assignment) rather than a top-priority emergency: there is no remote attack path, and turning the exposed hashes into working credentials requires offline cracking that succeeds only if the MQTT passwords are weak. The flaw is resolved in version 1.6.0; no public exploit code has been identified at time of analysis, and exploitation requires a local, low-privileged account on an affected image.
An input validation issue was addressed with improved input validation. This issue is fixed in iOS 26.7.2 and iPadOS 26.7.2, iOS 27.2 and iPadOS 27.2, macOS Golden Gate 27.2, macOS Sequoia 15.8.2, macOS Tahoe 26.7.2, visionOS 27.2. Processing web content may disclose sensitive information.
An out-of-bounds read in Apple's image-handling code lets a maliciously crafted image terminate the affected application when a user opens or renders it, affecting iOS and iPadOS, macOS and visionOS. Impact is confined to availability (C:N/I:N/A:H) and requires the victim to process the file (UI:R) with no authentication needed (PR:N); exploitation is limited to builds older than iOS/iPadOS 26.7.2 and 27.2, macOS Sequoia 15.8.2, macOS Tahoe 26.7.2, macOS Golden Gate 27.2 and visionOS 27.2. No public exploit code has been identified at time of analysis, the issue is not in CISA KEV, and EPSS is 0.60% (49th percentile); the realistic ceiling is denial of service, not data exposure or code execution.
Processing a maliciously crafted image on unpatched Apple devices can terminate the affected application, giving a denial-of-service condition on iOS/iPadOS, macOS, and visionOS. The flaw is an out-of-bounds read (CWE-125) that Apple addressed with improved bounds checking in iOS 26.7.2 and iPadOS 26.7.2, iOS 27.2 and iPadOS 27.2, macOS Golden Gate 27.2, macOS Sequoia 15.8.2, macOS Tahoe 26.7.2, and visionOS 27.2. The vector (AV:N/AC:L/PR:N/UI:R, C:N/I:N/A:H) means a remote attacker needs the victim to open or render the image, with no authentication; impact is limited to app termination and there is no public exploit identified at time of analysis, with EPSS at 0.58% (48th percentile) and CISA SSVC recording no exploitation and no automatable path.
An out-of-bounds write in Apple's image handling on iOS/iPadOS, macOS and visionOS can terminate the affected application when a user opens a maliciously crafted image, giving an availability-only denial-of-service outcome (CVSS 6.5, C:N/I:N/A:H). The attacker cannot trigger the fault directly: exploitation requires the victim to process the crafted file in an affected app, as reflected by the UI:R vector component, and macOS is affected across three branches (Sequoia before 15.8.2, Tahoe before 26.7.2, Golden Gate before 27.2) while iOS/iPadOS is affected before 26.7.2 and before 27.2 and visionOS before 27.2. Apple has released fixes for all listed platforms with improved bounds checking; no public exploit identified at time of analysis, EPSS is 0.58% (48th percentile), and CISA SSVC rates exploitation as none with partial technical impact, making this a moderate, not top-tier, priority.
An out-of-bounds write in Apple's image-processing path causes affected iOS, iPadOS and macOS systems to terminate an application when a maliciously crafted image is processed, giving an attacker a denial-of-service effect against a targeted app. The flaw affects iOS and iPadOS before 26.7.2 and before 27.2, plus macOS Sequoia before 15.8.2, macOS Tahoe before 26.7.2 and macOS Golden Gate before 27.2; no authentication or prior privilege is needed, but a victim must be induced to open the crafted image (UI:R), and the confirmed impact is availability only, with no disclosed code execution or data exposure. EPSS is low at 0.55% (46th percentile), SSVC records no known exploitation and no automatable path, and no public exploit identified at time of analysis; Apple has released fixed versions for every affected branch.
Out-of-bounds write in Apple iOS/iPadOS before 26.7.2 and 27.2, and visionOS before 27.2, can let a malicious or compromised app corrupt process memory or cause unexpected system termination. The vendor lists CVSS 6.5 with a network/user-interaction vector, but the app-based trigger and independent assessment indicate local execution with no privileges or user interaction, and the attacker must first get an app onto the device. No public exploit was identified at time of analysis (EPSS 0.55%), and Apple has released fixed versions.
An out-of-bounds write (CWE-787) in Apple's operating systems - iOS and iPadOS, macOS Sequoia, macOS Tahoe, macOS Golden Gate, and visionOS - allows a crafted or malicious app, or attacker-supplied content that a user interacts with, to trigger memory corruption and unexpected system termination. The impact is availability-focused: the supplied vector (PR:N/UI:R, C:N/I:N/A:H) and the independent assessment agree that confidentiality and integrity are not confirmed as affected, while system stability and process memory are. There is no KEV listing and no public exploit code identified at time of analysis, EPSS is low at 0.55% (46th percentile), and CISA SSVC rates exploitation as none and automatable as no, so this is routine patch-cycle work rather than an emergency; Apple fixed it with improved input validation across all current branches.
An out-of-bounds write (CWE-787) in Apple's image-processing code affects iOS and iPadOS before 26.7.2 and 27.2 and macOS Golden Gate before 27.2 and macOS Tahoe before 26.7.2; opening or otherwise causing a maliciously crafted image to be processed can crash the affected app, yielding a denial of service scored CVSS 6.5 (AV:N/AC:L/PR:N/UI:R/S:U/C:N/I:N/A:H). Exploitation requires user interaction - the victim must cause the image to be processed - so this is not a zero-click remote attack on default configurations, and the assessed impact is limited to unexpected app termination with no confidentiality or integrity loss. No public exploit is identified at time of analysis: SSVC records Exploitation: none and there is no KEV entry, though EPSS sits at 0.50% (42nd percentile); Apple has released fixed builds and the vendor cites improved bounds checking.
An out-of-bounds write in the macOS kernel's handling of mounted disk images lets a maliciously crafted image corrupt kernel memory or crash the system when a user mounts it. All macOS builds prior to Tahoe 26.7.2 and Golden Gate 27.2 are affected. The flaw is rated 6.5 (CVSS 3.1) by the vendor-facing sources, but exploitation requires the victim to mount the attacker-supplied image (UI:R) and confirmed impact is availability only, so this is a moderate-priority issue rather than an emergency: there is no CISA KEV listing, SSVC reports no known exploitation, EPSS is 0.46% (40th percentile), and no public exploit code was identified at time of analysis.
Apple macOS disk-image handling contains an out-of-bounds write that can corrupt kernel memory or terminate the system when a user mounts a maliciously crafted disk image. The flaw affects macOS versions before Golden Gate 27.2 and is fixed in 27.2; exploitation requires the victim to mount the crafted image (UI:R), needs no credentials (PR:N), is not automatable, and the confirmed impact is availability only, with no demonstrated code execution or data disclosure. No active exploitation is indicated and no public exploit was identified at time of analysis; EPSS is 0.44% (38th percentile), making this a moderate-priority issue rather than an emergency.
Kernel memory corruption in Apple macOS occurs when a user mounts a maliciously crafted disk image on versions earlier than macOS Golden Gate 27.2. The flaw is an out-of-bounds write (CWE-787) that can cause unexpected system termination or corrupt kernel memory; the assessed vector requires local user interaction to mount the image and confirms denial of service with limited integrity impact rather than demonstrated code execution. No public exploit has been identified at time of analysis, EPSS is 0.44% (38th percentile), and Apple has released the fix in macOS 27.2.
Denial of service in vLLM: any deployment serving the DeepSeek-OCR model family (deepseek-ai/DeepSeek-OCR, DeepSeek-OCR-2, or baidu/Unlimited-OCR) can be permanently killed by a single schema-valid chat request containing an ordinary half-black/half-white 1024x1024 PNG, whose pixel values cancel to exactly zero after normalization. That zero sum makes the image parser return no embedding, which trips a fatal encoder-output assertion inside the EngineCore worker; the API server raises EngineDeadError and exits, taking down every in-flight and future request until a manual restart. Affected versions run from 0.11.1 through before 0.32.0 on default configuration, the vendor advisory documents a verified end-to-end reproduction, but there is no CISA KEV listing and no public exploit code identified at time of analysis; the priority rating places remediation in the this-month window rather than emergency response.
An out-of-bounds write issue was addressed with improved bounds checking. This issue is fixed in macOS Golden Gate 27.2. Mounting a malicious disk image may cause unexpected system termination.
CPU and memory exhaustion in vLLM's /v1/responses endpoint affects versions 0.16.0 up to (but not including) 0.32.0, where an authenticated caller can inject a request-level Jinja chat template that bypasses the `trust_request_chat_template` policy gate. Only availability is impacted (no confidentiality or integrity loss), and exposure is limited to servers started with the documented default `trust_request_chat_template=False` and no explicit `--chat-template`, since that is the configuration in which the request-supplied `chat_template` key survives into the renderer. No public exploit code and no CISA KEV entry were identified at time of analysis, and the vendor's advisory rates the issue as Moderate.
Out-of-vocabulary token IDs slip past vLLM's prompt validation and can crash the inference engine, turning a single malformed request into a denial of service. Releases prior to 0.32.0 derive max_token_id from the tokenizer's vocabulary count rather than the largest valid index, so a pre-tokenized prompt carrying the ID equal to the vocabulary size is accepted and reaches the embedding lookup one row out of range, raising a CUDA device-side assert that poisons the context and terminates the engine. The crash is confined to models whose vocabulary is an exact multiple of the 64-row embedding padding size (Llama-2-family, vocab 32000) served at tensor-parallel size 1; the assessed CVSS 3.1 score is 6.5 (AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H), reflecting that low-privilege API access is needed and that availability is the only impacted property, and no public exploit was identified at time of analysis.
Denial of service in vLLM prior to 0.32.0 lets a remote client terminate the entire inference engine by submitting a single legal audio clip - a roughly 10-minute silent recording is enough, no crafted content required. When the server hosts microsoft/Phi-4-multimodal-instruct (or another model reusing its audio tower), the Phi-4 conformer encoder runs attention in eager mode by default, materializes a full attention probability matrix per layer for every 500-frame chunk, and retains it on the module, so GPU memory climbs monotonically within one request until a CUDA out-of-memory error kills EngineCore for all users; the advisory reports worker death in about six seconds and the API then returns EngineDeadError. No public exploit code was identified at time of analysis and the issue is not in CISA KEV; the authoritative assessment rates this a genuine but moderate-priority availability issue, rated high on availability only (C:N/I:N/A:H) and unauthenticated where the API is exposed, while deployments not serving this audio-capable model are unaffected.
Denial of service in vLLM 0.19.0 through 0.31.x lets a caller who holds a valid API credential hand the server a hostile Jinja2 chat template - either directly in the request when the operator has not set --chat-template, or indirectly through a model repository whose tokenizer_config.json carries one - and hang every chat-style endpoint on that API-server replica until the process is restarted. The render runs inside a sandboxed Jinja environment with no wall-clock timeout, no output-size cap and no loop or binop bound, on a render pool that defaults to a single non-interruptible worker, so a tiny non-terminating template wedges /v1/chat/completions, /v1/responses, /tokenize and the chat pooling/invocations paths for all tenants, while a string-multiplication variant instead OOM-kills the process. Impact is availability-only across tenants of one replica (CVSS:3.1 6.5, PR:L), Harmony/gpt_oss models skip the affected Jinja path, and no public exploit code was identified at time of analysis; the vendor shipped a fix in 0.32.0.
SOEM (Simple Open EtherCAT Master) 2.0.0 and earlier can be crashed by a malicious device on the same EtherCAT segment: an error response causes the library to clear an internal pointer while continuing to process, and a second message arriving before the timeout makes the master dereference that cleared pointer. The result is a NULL pointer dereference that takes down the master process, so impact is limited to availability (CVSS 6.5, AV:A/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H) with no code execution or data exposure. Exploitation requires the attacker to be a device on the local fieldbus, no public exploit code or confirmed active exploitation was identified at time of analysis, and the vendor advisory RRTL-260921-01 documents the issue.
Memory corruption in the rt-labs SOEM EtherCAT master library (version 2.0.0 and earlier) allows a malicious or compromised EtherCAT slave device to derail the master's device-configuration routine: by presenting more than EC_MAXFMMU FMMU configuration blocks, the slave forces out-of-bounds writes past a fixed-size array, and the resulting corruption of adjacent memory can disrupt subsequent master operation. The issue is scored CVSS 3.1 6.5 (AV:A/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H), so impact is confined to availability - no confidentiality or integrity loss and no confirmed code execution - and the main precondition is that an attacker can place or control a device on the adjacent EtherCAT segment. No public exploit code or confirmed active exploitation was identified at time of analysis; the vendor advisory RRTL-260921-01 is available.
An out-of-bounds read in the SOEM EtherCAT master library (version 2.0.0 and earlier) lets a malicious or compromised EtherCAT slave device send a response whose length field exceeds the actual payload, causing the master to over-read up to 64KB of its own memory and return that data to the attacker. Exploitation requires a foothold on the same EtherCAT network segment as the SOEM-based master - the flaw is not reachable over the public internet - and no authentication or user interaction is needed because EtherCAT has no native device authentication. Impact is confidentiality-only (memory disclosure); no integrity or availability effect is described, and no public exploit code or confirmed active exploitation was identified at time of analysis, though the vendor has published an advisory.
Catalog ingestion policy bypass in @backstage/plugin-catalog-backend lets an authenticated user holding the catalog.location.create permission register a URL whose raw string matches a trusted catalog.rules prefix (for example https://ghe.example.com/trusted/**) while the URL reader canonicalizes it to a location outside that prefix, so entity content from an untrusted source is ingested as if it came from a trusted one. Versions before 3.9.2 and 4.0.0 through 4.0.1 are affected, fixed in 3.9.2 (Backstage 1.54.10) and 4.0.2 (Backstage 1.56.1). Impact is scoped to the operator-configured ingestion policy boundary, downstream effects depend on deployment configuration, and no public exploit identified at time of analysis.
Stored cross-site scripting in the WP Chill Gallery PhotoBlocks plugin for WordPress, affecting versions through 1.3.7, allows a user holding a low-privilege WordPress account to persist malicious markup that executes in the browser of anyone who later views the affected gallery page. The CVSS 3.1 vector (AV:N/AC:L/PR:L/UI:R/S:C/C:L/I:L/A:L, score 6.5) reflects a network-reachable flaw that requires an authenticated low-privilege account and victim interaction, but whose scope change means the injected script can act beyond the vulnerable plugin component. There is no CISA KEV entry and no public exploit identified at time of analysis, and Patchstack's advisory is the underlying report; the vulnerable field or parameter is not described in the available sources.
Stored cross-site scripting in the bPlugins 'Typing Text - Give your texts many typing effects' WordPress plugin affects every release through 2.2.0, allowing an authenticated user with contributor- or author-level privileges to persist malicious markup that executes in another user's browser when the injected content is rendered. The assessed vector (CVSS 6.5, AV:N/AC:L/PR:L/UI:R/S:C/C:L/I:L/A:L) reflects that the attacker needs an existing account on the target site and that a second user must view the affected page before the payload fires, which limits the flaw to actors who already hold a foothold on the installation rather than remote unauthenticated intruders. No public exploit code or CISA KEV entry is associated with this issue at time of analysis, and the EPSS-style priority assigned by vuln.today places it in the current-month remediation queue rather than the emergency tier.
Stored cross-site scripting in the WordPress plugin Button Block (bPlugins) through version 1.2.7 lets an authenticated low-privilege user persist a script payload through the plugin's block content, which then executes in the browser of any higher-privileged user - typically an administrator or editor - who views or edits the affected page. The provider rates it CVSS 6.5 (CVSS:3.1/AV:N/AC:L/PR:L/UI:R/S:C/C:L/I:L/A:L), reflecting the requirement for a valid content-authoring account (PR:L) and for a victim to interact with the crafted content (UI:R). Exposure is meaningful mainly on multi-author sites that grant contributor or author roles to untrusted users; single-admin installations face substantially lower real-world risk, and no public exploit has been identified at the time of analysis.
Stored cross-site scripting in bPlugins Content Slider Block through version 3.2.1 lets an authenticated low-privileged user store a script payload that executes in a higher-privileged victim's browser when the victim views the affected block content. The CVSS 3.1 vector (AV:N/AC:L/PR:L/UI:R/S:C/C:L/I:L/A:L) scores 6.5, and exploitation depends on allowing untrusted users to author or configure blocks plus a victim with a higher-privileged session viewing that content. No public exploit was identified at time of analysis, and the exact unsanitized field is not described in the available sources; the vuln.today decision label is this_month.
NLnet Labs Routinator, an RPKI relying party validator, can be driven into memory exhaustion by a flood of concurrent requests to its HTTP JSON output endpoint, and can also abort with a stack overflow when a payload item carrying excessive metadata is discarded in a single operation. Only deployments that expose the HTTP or RTR server to untrusted networks are affected; the flaw is availability-only (no confidentiality or integrity impact), requires no authentication or user interaction, and no public exploit code was identified at the time of analysis. Vendor severity is 6.5 (CVSS:3.1/AV:A/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H), while the independent assessment rates the attack vector as network (AV:N) rather than adjacent, and the priority decision is this_month.
Stored cross-site scripting in the bPlugins LightBox Block WordPress plugin affects all versions through 1.2.0 and lets a low-privilege authenticated user (per the CVSS vector's PR:L) store script that runs when another user - typically an administrator - loads the affected page (UI:R). Rated CVSS 6.5 with scope change (S:C), it is a moderate, not emergency, priority under a this_month decision label, and no public exploit was identified at time of analysis.
Stored cross-site scripting in the WP Chill Gallery PhotoBlocks WordPress plugin, affecting versions through 1.3.7, lets an authenticated low-privileged user persist script payloads that run in another user's browser when the affected gallery content is rendered. CVSS 3.1 rates it 6.5 (AV:N/AC:L/PR:L/UI:R/S:C/C:L/I:L/A:L), with the scope change reflecting that impact lands in the viewer's session rather than in the plugin itself. No public exploit code has been identified at time of analysis, and no fixed version appears in the available sources, so the vendor's Patchstack advisory remains the authoritative reference for remediation status.
Stored cross-site scripting in the TMS Amelia WordPress booking plugin (all versions through 2.4.12) allows an attacker holding a low-privilege authenticated account on the target site to persist script payloads that execute when another user - typically a higher-privileged operator or administrator - views the affected page. Rated CVSS 6.5 (CWE-79) with scope change (S:C), the flaw can read and alter limited data in the victim's browser context; exploitation is constrained by the need for an account (PR:L) and victim interaction (UI:R). Patchstack is the reporting source, no public exploit code or in-the-wild activity is identified in the available data, and the priority label is 'this_month'.
Integrity bypass in the zip2 Rust crate (0.6.0 through versions below 9.0.0) lets an attacker who can modify or supply a ZIP archive deliver altered or attacker-chosen plaintext from an AES-encrypted entry that is read through the password-bearing APIs, with no authentication error raised. The HMAC-SHA1-80 tag in src/aes.rs is only checked when the reader's byte counter reaches zero, so a crafted entry that hits true end-of-file while the header-declared compressed_size still leaves data remaining - or an entry with zero payload - returns a clean end of stream and skips the tag comparison entirely; because AES-CTR is malleable, the modification needs no knowledge of the password. Scored CVSS 6.5 (AV:N/AC:L/PR:N/UI:R/S:U/C:N/I:H/A:N), it is unauthenticated from the attacker's side but requires the victim to open the crafted archive with a valid password through by_index_decrypt, by_name_decrypt, by_path_decrypt, or by_index_with_options; impact is integrity-only, with AE-2 entries additionally losing their by-design-disabled CRC-32 check so no integrity verification of any kind remains. A public PoC reference exists in the form of the advisory's test code; it is not verified as working exploit code outside that context, and no active exploitation is indicated. Fixed in 9.0.0.
A null-pointer dereference in the Zephyr Bluetooth Classic Hands-Free Profile Audio Gateway role can crash or hang the Bluetooth host stack when an RFCOMM connect fails after an SDP discovery has already linked the AG object into the SDP client's pending request list. Builds of Zephyr 4.4.0 through 4.4.2 with CONFIG_BT_CLASSIC and CONFIG_BT_HFP_AG enabled and acting as an Audio Gateway are affected; a nearby BR/EDR device advertising an out-of-range RFCOMM channel (outside 0x01..0x1e) in its SDP Hands-Free record, which requires no pairing to read, deterministically forces the failure path. Impact is availability only - crash or hang of the Bluetooth stack, with no attacker-controlled write - and no public exploit code has been identified at time of analysis; the vendor fix is merged on main and v4.4-branch with the release carrying it forecast as 4.5.0.
DHCPv6 option parsing in Arm Keil MDK-Middleware 8.0.0 through 8.3.0 can over-read memory and unintentionally retransmit data, allowing an attacker on the same link-local network segment to obtain limited information and disrupt availability when IPv6/DHCPv6 is enabled. No authentication or user interaction is required, but impact is limited to low confidentiality loss and low availability disruption, with no integrity impact or code-execution claim in the available sources. No public exploit was identified at time of analysis, and this is a moderate-priority issue rather than an emergency.
Out-of-bounds reads and writes in the FTP Server RX buffer of Arm Keil MDK-Middleware 8.3.0 let network-reachable input drive memory accesses beyond the intended receive buffer, with the assessed vector (CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:L/A:L, 6.5) indicating low complexity, no user interaction, no confidentiality loss and only partial integrity and availability impact. Only firmware and devices that actually build in and expose the optional MDK-Middleware FTP Server component are exposed, and the sources do not state whether the out-of-bounds condition is reached before or after FTP authentication, so an FTP login may still be a limiting prerequisite. No public exploit code has been identified at time of analysis, there is no CISA KEV entry, and SSVC records exploitation as none with partial technical impact while marking the flaw automatable.
When cross-site (xsite) TLS is enabled, the Infinispan Operator for Red Hat Data Grid 8 launches the Gossip Router container with the keystore and truststore passwords as plain-text command-line arguments, so any principal holding `get pods` in the namespace - without needing `get secrets` - can read those credentials directly from the Pod spec. The same values are exposed to node-level processes through `/proc/<pid>/cmdline` and are frequently captured in audit and event logs. Rated 6.5 (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N), this is a moderate-priority information-disclosure issue that only manifests in a non-default operator-driven configuration; no public exploit code and no CISA KEV entry were identified at time of analysis.
Mattermost 11.7.x through 11.10.x servers can be driven into memory exhaustion by an authenticated user who uploads a crafted PDF whose compressed content stream expands into a very large amount of extracted text, causing a denial of service; the CVSS vector is availability-only (6.5, PR:L). The priority signal is this_month, consistent with scheduled remediation rather than an emergency, and no public exploit has been identified at the time of analysis.
Unbounded memory allocation in the Mattermost WebSocket msgpack decoder lets any authenticated user exhaust server memory by sending a single crafted binary WebSocket frame that declares a byte-string payload length of up to 4,294,967,295 bytes. All four maintained branches are affected: 11.7.x through 11.7.10, 11.8.x through 11.8.5, 11.9.x through 11.9.1 and 11.10.x through 11.10.1; the attacker needs only a standard account plus a normal WebSocket connection, with no administrator rights and no victim interaction (CVSS 3.1 6.5, AV:N/AC:L/PR:L/UI:N, availability-only). No public exploit code has been identified at time of analysis and there is no CISA KEV entry, so the vuln.today priority label is this_month; one input tag classifies this as information disclosure, but the assessed impact is exclusively to availability.
Denial of service in the Mattermost Agents plugin: when the plugin summarizes a Zoom chat transcription that contains very short continuation lines, it slices those lines without first validating their length, and the resulting error crashes the plugin. Mattermost versions at or below the 11.9, 11.1.9, 11.5.8 and 11.10.7.0 branch levels are affected, and an attacker needs an authenticated account (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H) plus a deployment where the Agents plugin and its Zoom transcription summarization are actually in use; impact is limited to availability of the plugin rather than the whole server. No public exploit code has been identified at time of analysis and the issue is not listed in CISA KEV; severity is moderate (6.5) and vendor priority is this-month rather than emergency.
Mautic's Stats API exposes internal application tables to any authenticated API user, including an account whose role grants no permissions, because permission checks for several tables were missing or an intended administrator-only restriction was never enforced. The flaw affects Mautic releases before 5.2.12, 6.0.10, and 7.2.2 and can disclose audit-log actions and change details, stored visitor IP and geolocation data, and tracked redirect URLs with channel mappings. Exploitation requires a valid API account and an enabled REST API; no public exploit identified at time of analysis, and vendor-released fixed versions are available.
Path traversal in Mautic's Asset API lets an authenticated API user with asset-creation permission register a file outside the asset upload directory and publish it so its public download URL exposes arbitrary files readable by the web server, including Mautic configuration secrets. The affected composer/mautic/core branches are fixed in 7.2.2, 6.0.10, and 5.2.12; no public exploit was identified at the time of analysis.
Authenticated path traversal in Logsign SIEM 6.4.120 through 6.4.121 allows a logged-in attacker to read files outside the intended directory, exposing potentially sensitive data with high confidentiality impact but no integrity or availability effect. The issue is scored CVSS 6.5 (AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N) and affects builds from 6.4.120 before 6.4.122. No public exploit or CISA KEV entry was identified at time of analysis, and the specific vulnerable endpoint and pathname parameter are not described in the available sources.
Stored cross-site scripting in the Crocoblock JetSmartFilters WordPress plugin, affecting versions through 3.8.6.1, lets an authenticated low-privileged user persist script content that later executes in another user's browser when the affected page is loaded. The CVSS 3.1 score of 6.5 includes a scope change (S:C), reflecting that the injected script runs in the victim's browser session rather than in the server's security context. No CISA KEV entry and no public exploit code were identified at the time of analysis, and the issue is prioritized for patching this month rather than treated as an emergency.
Stored cross-site scripting (CWE-79) in the bPlugins 'Advanced Post Block - Post Grid for WordPress block editor' plugin, affecting every release through 2.5.1, allows a user holding a low-privileged authenticated account to persist script-bearing input that executes in another user's browser when that content is rendered. The assessed vector (CVSS:3.1/AV:N/AC:L/PR:L/UI:R/S:C/C:L/I:L/A:N) makes this a moderate, not urgent, priority: an attacker needs a valid account on the target site plus a victim who views the affected page, so exposure is concentrated on multi-author sites that grant contributor or author-level access to untrusted people. No public exploit code and no confirmed active exploitation were identified at time of analysis, and the available intelligence lists no fixed version beyond 2.5.1.
Stored cross-site scripting in the bPlugins 'Panorama Viewer - 360 Degree Image + Video Viewer' WordPress plugin, affecting all versions through 1.8.3, lets a low-privileged authenticated user plant a script payload that executes in the browser of a second user who later views the affected panorama content or admin screen. Rated CVSS 6.5 (CVSS:3.1/AV:N/AC:L/PR:L/UI:R/S:C/C:L/I:L/A:L), it is not unauthenticated and not self-triggering: the attacker needs a WordPress account able to create or edit panorama content, and a victim - administrator or site visitor - must render the page holding the stored payload. No public exploit code and no confirmed active exploitation were identified at time of analysis, and the vendor's Patchstack advisory is the reference point for a fix.
Stored cross-site scripting in the bPlugins 'B Slider' WordPress plugin (versions through 2.2.2) allows an authenticated low-privileged user to persist script payloads that execute in the browser of anyone who later views the rendered slider, carrying a CVSS 3.1 base score of 6.5 (AV:N/AC:L/PR:L/UI:R/S:C/C:L/I:L/A:L). The flaw is only reachable where the plugin is installed and accounts with content-contribution privileges (contributor/author level) can submit or edit slider content, and the payload fires only when a victim with an active browser session loads the affected page - typically an administrator or editor, which is where the scope-changed impact becomes meaningful. No CISA KEV entry or public exploit code was identified at time of analysis; the vuln.today priority label is 'this_month' and the entry was reported through Patchstack.