Privilege escalation in IBM Financial Transaction Manager (FTM) for Red Hat OpenShift lets a remote authenticated attacker who already holds a privileged account exceed the permissions assigned to that account, with high impact across confidentiality, integrity and availability. IBM rates it 9.1 (CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:C/C:H/I:H/A:H), a score driven mainly by the Scope-Change flag; our independent assessment preserves the full CIA impact but treats scope as unchanged, because the high starting-privilege requirement narrows the realistic attacker population to insiders or holders of compromised privileged credentials, and the deployment footprint is limited to organizations running FTM on OpenShift. No public exploit code has been identified at time of analysis and the issue is not listed in CISA KEV; IBM has published a vendor patch advisory.
Authenticated low-privileged members of WuzhiCMS 5.0.0 can convert a routine thumbnail upload into full remote code execution, because the member article thumbnail handler at /index.php?m=member&f=article&v=thumbUpload saves the client-supplied file extension verbatim and writes the file into the web-accessible uploadfile/ directory, where a PHP-enabled web server executes it. The CVSS 3.1 vector (AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H, 8.8) confirms the attacker needs only a valid low-privilege account and no user interaction, while SSVC rates technical impact as total; publicly available exploit code exists, so the only real barrier is obtaining a member login, which is trivial where self-registration is open and a meaningful obstacle where account creation is restricted. There is no CISA KEV entry indicating confirmed active exploitation, and the risk is best characterized as a consistently high-priority, easily automated-adjacent upload flaw rather than a confirmed active campaign.
Authenticated remote code execution in openEQUELLA releases before 2026.1.0 lets any logged-in non-guest user execute arbitrary code on the server by sending a crafted serialized Java object to the HTTP invoker endpoint at /invoker/*. The flaw is a denylist bypass: the class-name filter in PluginAwareObjectInputStream is evaded by wrapping a malicious payload inside a java.security.SignedObject, whose inner stream is handled by a separate ObjectInputStream that never applies the denylist, ultimately routing into a JNDI lookup sink. No public exploit identified at time of analysis and the vulnerability is not in CISA KEV; the practical barrier is only the low-privilege account requirement (PR:L per the assessed CVSS:3.1 vector AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H, scored 8.7 under CVSS 4.0), which is trivial to satisfy on instances that permit self-registration or broad SSO/LDAP provisioning.
CMSimple 5.24 ships with CSRF protection disabled by default, so an unauthenticated attacker can forge a content-save request that a currently logged-in administrator's browser submits (CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H, 8.8), storing a text payload containing a scripting marker verbatim in content/content.php. Because evaluate_cmsimple_scripting() in functions.php subsequently runs that marker through PHP eval() on every page view, the result is persistent remote code execution on the web server for all visitors, including unauthenticated ones. There is no public exploit identified at time of analysis and the issue is not in CISA KEV; EPSS is 0.33% (26th percentile), and realistic risk is limited by the requirement to socially engineer a privileged, actively-authenticated administrator who holds a valid session cookie.
Improper enforcement of the behavioral workflow in NVIDIA Infrastructure Controller for Linux (versions 0 through 1.9) lets an authenticated low-privilege user on the management plane drive the product's API with out-of-order or state-skipping requests that defeat intended workflow checks, with possible data tampering, denial of service, and information disclosure. Exploitation is not remote-unauthenticated: it requires a valid low-privilege account plus the ability to craft non-sequential workflow calls, and SSVC records Exploitation as 'none' and Automatable as 'no', so there is no public exploit identified at time of analysis. EPSS is low at 0.31% (24th percentile), and our independent assessment rates the practical severity below the vendor's 8.8 because the assessed vector is AC:H with only low confidentiality, integrity, and availability impact.
Hard-coded credentials in NVIDIA Infrastructure Controller for Linux (versions 0 through 1.9) let an unauthenticated attacker who can reach the device from an adjacent network segment log in with embedded, static credentials, resulting in high confidentiality, integrity, and availability impact (CVSS 3.1 8.8, AV:A/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H). Because the password is static across installations and recoverable by anyone who can inspect the binary or firmware image, exploitation requires no user interaction and no prior authentication, but it does require the attacker to reside on the same broadcast or management network segment rather than the open internet. No public exploit code has been identified at time of analysis, there is no CISA KEV listing, and EPSS is low at 0.23% (14th percentile), with CISA SSVC rating exploitation as "none" and the attack as not automatable.
Horizontal privilege escalation in the x-ui web panel 0.3.2 allows any authenticated panel user to rewrite another user's inbound proxy configuration - remark, port, protocol, settings, enabled state, expiry time and traffic quota - by submitting an inbound-update request that carries the victim's resource identifier, because the update path never verifies that the referenced inbound belongs to the requesting session user (CWE-284). The flaw is network-reachable and requires no victim interaction, with a CVSS 3.1 vector of AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H (8.8); an independent assessment retains High severity and confirms the integrity impact (cross-tenant configuration tampering) while narrowing confidentiality and availability to Low, since the tampering is confined to proxy resource data rather than granting panel or host compromise. Exploitation is limited to multi-tenant deployments where more than one account holds inbound resources; no public exploit code has been identified at time of analysis and EPSS is low (0.20%, ~10th percentile), so mass scanning is unlikely, though the flaw is trivially reachable by any legitimate panel account.
IBM Financial Transaction Manager (FTM) for Red Hat OpenShift contains a command execution flaw stemming from functionality being pulled in from an untrusted control sphere (CWE-829), so an attacker who already holds a low-privilege foothold in the OpenShift environment hosting the product can influence the untrusted source and cause arbitrary commands to run within the FTM process context. The attack is local (AV:L, PR:L, UI:N) - there is no unauthenticated or internet-facing path to the flaw - which limits exposure to environments that actually run the affected FTM build. No public exploit code has been identified at time of analysis and the flaw is not listed in CISA KEV; the vendor has published a patch, and the vendor base score of 8.8 (CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H) reflects a scope-change assumption that independent assessment does not confirm, so real-world severity should be treated as high-but-localized rather than remotely critical.
Local privilege escalation in Qualcomm Snapdragon platforms allows a low-privilege on-device process or app to cross a trust boundary and gain system-level control by abusing insecurely handled temporary files. The issue is tracked as CVE-2026-25265 with a vendor/assessed CVSS 3.1 score of 8.8 (AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H), meaning exploitation requires local access and low privileges but no user interaction, and results in high confidentiality, integrity, and availability impact. No public exploit code or confirmed active exploitation was identified at time of analysis, and affected Snapdragon chipset/firmware versions are not enumerated in the input; confirm scope via the Qualcomm June 2026 product security bulletin.
Privilege escalation in Qualcomm Snapdragon devices allows a local low-privilege attacker to cross security boundaries and gain full control over confidentiality, integrity, and availability (CVSS 8.8). The flaw stems from a weak configuration during package extraction (CWE-427) that lets an attacker-controlled element be resolved or loaded with higher privilege; exploitation requires prior local code execution but no user interaction. No public exploit code or active exploitation has been identified at time of analysis.
Predictable digital-product download tokens in the Concrete CMS Community Store package before 2.7.8 allow unauthenticated remote attackers to download files belonging to other customers. The token protecting each download URL is an MD5 hash of the order's creation timestamp - a value with only one-second resolution - so an attacker who can enumerate valid order and file identifiers can brute-force the token and retrieve purchased digital goods. The flaw is remotely exploitable without authentication or user interaction (CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N, vendor-assigned CVSS 4.0 score 8.7) and affects any Concrete CMS site selling downloadable products with this package; no public exploit code was identified at time of analysis, though the deterministic token construction makes weaponization trivial once identifiers are known.
Unauthenticated remote attackers who can reach the Vector `logstash` source listener (default 0.0.0.0:5044) can force a multi-gigabyte memory allocation and terminate the process by sending a single minimal Lumberjack frame that merely declares a huge compressed-frame length, without transmitting those bytes. Deployments running the logstash source in Vector 0.15.0 through 0.56.x are affected; the issue is fixed in 0.57.0 by capping declared and decompressed frame sizes. Impact is availability only (crash or host OOM kill that halts log ingestion for every tenant on a shared pipeline) with no code execution or data exposure, and no public exploit code identified at time of analysis.
Denial of service in MikroTik RouterOS: an on-path attacker able to inject a BGP UPDATE into an established session can crash the device's BGP routing service with a single malformed MP_REACH_NLRI UPDATE whose labelled-VPN NLRI carries a prefix-length below the minimum valid value. Affected are RouterOS builds before 7.25beta4 that run the routing/BGP service with the labelled-VPN VPNv4 or VPNv6 address family enabled; the malformed route passes NLRI validation, describes an address portion of negative length, terminates the session without a NOTIFICATION, and can be resent to hold down the BGP plane indefinitely. The exploit requires a peer or man-in-the-middle position on the peering link rather than arbitrary off-path internet access, no public exploit code was identified at time of analysis, and the vendor has released a fix in the 7.25beta4 development build.
Stored cross-site scripting in MCMS 6.1.1 through 6.2.1 allows an authenticated user with article authoring or submission privileges to inject JavaScript into the contentDetails field, which is excluded from the global XSS filter; a victim, such as an administrator, editor, or site visitor, must view the stored article for the payload to execute (UI:R). Publicly available exploit code exists, and the EPSS score is low at 0.19% (9th percentile), with no confirmed active exploitation from the provided KEV data. The independent assessment rates this as moderate priority with low confidentiality and integrity impact, not the high impact suggested by the original 8.7 feed score.
Server-side request forgery in Adobe Experience Manager Forms JEE lets an attacker who already holds high-privileged, admin-level application access coerce the AEM Forms JEE server into issuing requests to internal-only resources, achieving privilege escalation and a scope change beyond the original AEM security perimeter. Affected deployments are AEM 6.5 Forms JEE and AEM 6.5 LTS Forms JEE, reachable over the network (AV:N) with no end-user interaction (UI:N) but gated behind the PR:H privileged-authentication requirement, so practical risk concentrates on malicious insiders or attackers who have already compromised an administrative account. Based on the supplied intelligence there is no CISA KEV listing and no public exploit code identified at time of analysis; the high EPSS-style urgency implied by the CVSS 8.7 base score reflects the scope change and confidentiality/integrity impact rather than demonstrated exploitation.
Remote, unauthenticated attackers who can reach the DDS/RTPS data bus of an application built on RTI Connext Professional Core Libraries can trigger abnormal middleware behavior and degrade or halt that application's availability; the root cause is an incorrect calculation (CWE-682) that can be exploited by sending crafted DDS/RTPS traffic with no authentication, user interaction, or non-default configuration. All shipping branches are covered by the advisory: 7.4.0 through 7.7.0.0, 7.0.0 through 7.3.1.5, the 6.1.x, 6.0.x, 5.3.x and 5.2.x lines, and 4.1x through 5.0.x, with fixes available at 7.7.0.1, 7.3.1.6, and the 5.1.x legacy line. The impact is availability-only per our assessment (CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H), and no public exploit code or confirmed active exploitation was identified at time of analysis; the vendor's own CVSS 4.0 score of 8.7 reflects the same unauthenticated, network-reachable, high-availability-impact profile.
Credentialed cross-origin reads and namespace takeover were possible against the hosted open-vsx.org registry because its CDN/edge layer reflected the requesting Origin back in Access-Control-Allow-Origin while simultaneously emitting Access-Control-Allow-Credentials: true on authenticated /user/ endpoints. A page on any origin could then issue credentialed requests in a logged-in user's browser and read the responses, exposing /user (login name, avatar, homepage, tokens URL), /user/tokens, /user/namespaces, /user/extensions, /user/search/{name}, /user/namespace/{name}/members and - critically - /user/csrf; with the CSRF token readable, the write-endpoint protection collapsed, and chaining /user/csrf with /user/token/create let the attacker page mint and exfiltrate a personal access token carrying publish and delete rights over the victim's namespaces. Attackers need no authentication of their own (CVSS 4.0 PR:N) and only passive victim interaction - the victim must be logged in to open-vsx.org and visit the attacker's page - while no public exploit code has been identified at time of analysis.
Aureus ERP installations running the Webkul Chatter plugin before version 1.5.0 expose every message, activity, and note record to any authenticated user through an insecure direct object reference in the ChatterPanel Livewire component. Because the component resolved records with an unscoped Message::find($id) call instead of constraining the lookup to the parent record, an attacker holding nothing more than a low-privileged login (CVSS PR:L, no user interaction, no non-default configuration) can supply sequential message IDs to read, edit, delete, or pin messages belonging to other departments or companies and enumerate all notes in the system - an impact that is severe in multi-company or multi-department deployments where cross-tenant isolation is assumed. The assessed risk is a genuine, well-substantiated IDOR with high confidentiality and integrity impact (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:N; vendor-published CVSS 4.0 base 8.6), and no public exploit code was identified at time of analysis, although a public technical writeup of the issue is referenced. The vendor has released a fix in version 1.5.0.
Improper input validation in Adobe Campaign Classic allows a low-privileged, authenticated attacker with network reachability to achieve arbitrary code execution in the context of the current user, with the impact scoped beyond the vulnerable component (CVSS 3.1 base 8.5, AV:N/AC:H/PR:L/UI:N/S:C/C:H/I:H/A:H). No user interaction is required, but Adobe explicitly notes that exploitability depends on conditions beyond the attacker's control, reflected as AC:H - meaning successful exploitation is gated by specific environmental, timing, or deployment state rather than being a reliably reproducible remote attack against default installs. There is no public exploit identified at time of analysis and the issue is not confirmed as actively exploited (no CISA KEV listing), and the specific vulnerable feature or endpoint is not named in the advisory summary.
Vaultwarden through 1.37.3 permits authenticated users who hold a revoked or not-yet-confirmed organization membership to retain read, write, delete, and attachment access to organization ciphers, because three cipher access-restriction queries omit a membership status filter. Exploitation requires a Vaultwarden account with a prior (revoked) or pending membership linkage to the target organization and only affects deployments that use collection- or group-based cipher sharing, so a wholly external user with no org relationship cannot exploit it; no public exploit code was identified at time of analysis, and the issue is fixed upstream via PR #7554 (upgrade past 1.37.3).
Missing authentication in the ManageEngine ADSelfService Plus REST API (all builds before 7001) lets remote, unauthenticated attackers reach API functionality that should require a valid session or credential, as classified under CWE-306. The assessed vector is CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:H (8.6), meaning no credentials and no user interaction are needed, with the highest rated component being availability impact and more limited confidentiality/integrity effects - the exact reachable endpoints are not disclosed in the vendor advisory. No public exploit code and no confirmed active exploitation were identified at time of analysis, so EPSS/KEV signals cannot be used to prioritize this; risk is driven by internet-exposed instances of a pre-7001 build, and patching to build 7001 removes the exposure entirely.
Local privilege escalation to root in Netdata versions prior to 2.10.4: the setuid-root ndsudo helper command fail2ban-client-status-socket accepted a caller-supplied --socket_path argument from the low-privileged netdata service account, letting that account point root's fail2ban-client at an attacker-controlled UNIX socket. Because fail2ban-client's CSocket.receive() deserializes the socket response with pickle.loads(), a crafted reply yields attacker-controlled code execution as root on any host where fail2ban-client is installed and the setuid ndsudo helper is present (default in affected packaged builds). The vector is strictly local (CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H), so a prior foothold as the netdata account is mandatory, and hosts without fail2ban-client are not exploitable through this path; no public exploit code has been identified at time of analysis.
Local code execution on Windows hosts running the Apache Lounge distribution of Apache HTTP Server (builds before Apache 2.4.68-260920 Win64) is possible because the build hardcodes the path to openssl.cnf, allowing a local attacker who can write to that path to plant a malicious OpenSSL configuration that the server loads when TLS is initialized. The vulnerability is rated CVSS 8.4 (AV:L/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H) and classified CWE-732, meaning the weakness is an incorrect permission assignment for a critical resource; exploitation requires the attacker to already have local interactive or code-execution access on the host and the hardcoded path to be writable at the attacker's privilege level, with successful exploitation affecting confidentiality, integrity, and availability at high impact. No public exploit identified at time of analysis and no CISA KEV listing applies, and the issue is specific to the third-party Apache Lounge Windows build rather than the upstream Apache Software Foundation source or Linux distributions.
Heap use-after-free in the XS backend of Net::IDN::Punycode 2.301 through 2.589 allows an attacker-supplied punycode label that decodes to enough code points above U+FFFF to corrupt memory in the decoding Perl process, granting both read and write access to freed heap memory (CWE-416, CVSS 8.4). The defect was introduced by the 2.301 fix for CVE-2016-15059 and is confined to the XS implementation; deployments using the pure-Perl backend are not affected, and no public exploit code has been identified at time of analysis. Reachability depends on whether the consuming application feeds untrusted punycode into decode_punycode (for example IDN hostnames or internationalized email addresses), but the vulnerable call itself requires no authentication (CVSS PR:N).
MCP Atlassian servers before 0.22.0 write their plaintext OAuth fallback token file using permissions inherited from the process umask, so on a shared host with a permissive or default umask another local account can read the file and reuse the stored refresh and access tokens for Atlassian (Confluence/Jira) access. The flaw is a local, multi-user exposure rather than a remote one: the attacker must already hold a separate account on the same machine and the deployment must actually use the plaintext fallback storage path, which is why the independent assessment scores it CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:L/A:N instead of the vendor's AV:N rating. Because the leaked refresh token keeps working even after access-token rotation, credential rotation alone does not close the exposure. No public exploit code has been identified at time of analysis, and the issue is fixed in the vendor-released version 0.22.0.
Improper TLS certificate validation in NVIDIA Infrastructure Controller for Linux (versions 0 through 1.9) allows an attacker who already holds a low-privilege authenticated session and sits on the same adjacent network segment to man-in-the-middle controller TLS connections, leading to information disclosure, data tampering, and denial of service. The flaw is rated CVSS 7.6 (AV:A/AC:L/PR:L/UI:N, C:H/I:H/A:L) and there is no public exploit code identified at time of analysis, no CISA KEV listing, and an EPSS score of only 0.13% (3rd percentile), consistent with a genuine but moderate, non-urgent risk. Exploitation is constrained by the requirements for adjacency (not internet-reachable) and prior low-privilege authentication, so the realistic attacker set is narrow.
Heap-based memory corruption in RTI Connext Professional Core Libraries (7.4.0 before 7.7.0.1, 7.0.0 before 7.3.1.6, and the 6.1, 6.0, 5.3, and 5.2.3 branches) lets unauthenticated remote attackers who can reach a DDS domain corrupt heap memory by sending crafted RTPS/DDS traffic, yielding high integrity and availability impact with no confidentiality exposure. No public exploit has been identified at time of analysis and the issue is not confirmed as actively exploited (CISA KEV), though the CVSS 4.0 AT:P metric indicates exploitation depends on a specific attack precondition such as a particular message state or sequence. Risk is concentrated in environments where DDS traffic is reachable from untrusted network positions rather than in isolated OT, robotics, or simulation segments.
Denial of service in Erlang/OTP's ASN.1 OBJECT IDENTIFIER decoders allows a remote, unauthenticated attacker to stall TLS-processing services by sending a crafted certificate during the handshake. Erlang/OTP releases from OTP 17.0 up to the fixed versions OTP 27.3.4.18, OTP 28.5.0.7 and OTP 29.1.1 are affected whenever the service parses a peer's X.509 certificate - the default for TLS clients (which always parse the server certificate) and for mutual-TLS servers - because the vulnerable BER OID decoder runs before signature and trust-chain verification, so an untrusted self-signed certificate is sufficient to trigger it. A single crafted OID with one very large arc (roughly 262 KB of continuation bytes) burns about 13 seconds of CPU, so sustained denial of service requires repeated connections; no public exploit code identified at time of analysis, and the impact is availability only (vendor CVSS 8.2; independently assessed as CVSS 3.1 AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H).
Unauthenticated local file inclusion in the HUSKY - Products Filter for WooCommerce Professional WordPress plugin (all versions up to and including 1.4.4) allows any site visitor to pass an attacker-controlled value through the plugin's 'shortcode' parameter and force the server to include and execute arbitrary .php files, enabling access-control bypass, disclosure of sensitive data, and - where an attacker-controllable .php file already exists on the server (upload, log poisoning, or session files) - full remote code execution. The only access control is a nonce check against 'woof_front_nonce', which the plugin emits into inline JavaScript on every front-end page, so the nonce is freely obtainable without authentication; the assessed CVSS vector is CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:H (8.1, High) with High confidentiality, integrity, and availability impact. No public exploit code has been identified at time of analysis and there is no CISA KEV entry, but the preconditions are met on default configurations of any WooCommerce site running the affected plugin, making this a genuine high-priority issue rather than a high-CVSS paper tiger.
XML injection in NVIDIA Infrastructure Controller for Linux (affected versions 0 through 1.9) allows an attacker who already holds a low-privilege authenticated session to corrupt data and degrade availability by feeding crafted input into the product's XML handling. The flaw is network-reachable (AV:N), needs no user interaction (UI:N), and leaves confidentiality untouched (C:N), with the full impact landing on integrity and availability (I:H/A:H); the practical exposure is narrowed because valid credentials are mandatory, the target is a specialized NVIDIA datacenter infrastructure component, and SSVC rates the issue non-automatable with only partial technical impact. No public exploit code has been identified at time of analysis, EPSS places exploitation likelihood at roughly 0.5% (43rd percentile), and there is no CISA KEV listing - this is a genuine medium-to-high priority fix, not an emergency.
Remote code execution via ESQL command injection affects IBM Financial Transaction Manager (FTM) for Red Hat OpenShift, where externally supplied input is combined into ESQL statements without neutralization of special elements. Per IBM PSIRT scoring (CVSS 8.1, AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:H), an unauthenticated remote attacker who can reach a vulnerable FTM message flow that builds ESQL commands from attacker-controlled data can achieve full confidentiality, integrity, and availability impact within the FTM runtime and its connected resources. The high attack-complexity rating reflects a non-default precondition - typically a specific flow or configuration that routes untrusted data into ESQL - so exposure is limited to environments running FTM on OpenShift with such a flow. No public exploit code identified at time of analysis and the issue is not listed in CISA KEV; EPSS data was not provided.
Replay of a captured Tempo subscription credential in ZenHive mpp 0.14.0 through 0.16.1 lets an attacker who already holds a validly signed Tempo KeyAuthorization repeatedly charge the same payer's wallet. The signed authorization is bound only to chain id, key type, key id, expiry, limits, and scopes - never to the challenge that prompted it - so it verifies against every challenge the server issues for the same subscription terms, and because Subscription.activate/4 deduplicates activations solely by challenge id, a fresh challenge yields a new dedup key and the settlement transaction is rebuilt and broadcast again. No public exploit code has been identified at time of analysis and there is no CISA KEV entry, but the flaw class is a straightforward capture-replay, exploitation requires only that the attacker obtain one signed credential (attack requirements: present, AC:H), and the independent assessment rates integrity impact High on a CVSS:3.1 base of 8.2.
Authenticated shortcode injection in the WP Ultimate Review WordPress plugin (Roxnor) allows any user with subscriber-level access or higher to run arbitrary registered shortcodes on sites running version 2.4.2 or earlier. The attacker passes a crafted value to a plugin action that feeds it into do_shortcode() without validation (CWE-94), yielding high confidentiality and integrity impact per the assessed CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:N (8.1). No public exploit code was identified at time of analysis and there is no confirmed active exploitation (CISA KEV), but on sites with open registration the subscriber prerequisite is trivial to satisfy, and the real-world severity is bounded by which powerful shortcodes are registered by this or other installed plugins and themes - arbitrary operating-system command execution is not guaranteed.
Stored cross-site scripting in Adobe Experience Manager 6.5 Forms JEE and 6.5 LTS Forms JEE lets a high-privileged, authenticated attacker plant malicious JavaScript inside form fields that executes in a victim's browser when the poisoned form page is rendered, with a scope change that can extend impact to the victim's session or account. The CVSS vector confirms the prerequisites: PR:H means the attacker must already hold an elevated AEM account able to write form content, and UI:R means a victim must actively browse to the affected page, so realistic exposure is limited to malicious insiders or attackers who have already compromised high-privilege AEM credentials. No public exploit code has been identified at time of analysis and the issue is not listed in CISA KEV; the 8.1 score is driven by the high confidentiality and integrity impact on the victim's session rather than by an easy unauthenticated path, making this a genuine but constrained-priority issue rather than a drop-everything emergency.
Authenticated viewer accounts in Frigate NVR can abuse the WebSocket message handler to invoke admin-only control topics, causing service restarts and disabling security monitoring functions across cameras. The flaw affects all releases prior to 0.17.2 and stems from the nginx authentication subrequest proving identity without conveying role information, so frigate/comms/ws.py forwards client-supplied topics to the dispatcher without any role check (CWE-862, missing authorization). Exploitation requires authentication to be enabled plus valid viewer-level credentials, meaning the realistic threat is an insider or the holder of a leaked or shared low-privilege account rather than an anonymous internet attacker; no public exploit code has been identified at time of analysis, and the vendor-released patch is version 0.17.2.
Code execution, information disclosure, and data tampering are possible in NVIDIA NeMo Speech versions 0.0 through 2.9 due to unsafe deserialization of untrusted data. Exploitation requires the victim to load attacker-crafted serialized data (dataset, manifest, or model checkpoint) on their own machine, making this a local, user-interaction-dependent flaw (AV:L/UI:R) that runs with the victim's privileges. No public exploit code or confirmed active exploitation (CISA KEV) has been identified at time of analysis.
Deserialization of untrusted data in the NVIDIA NeMo Speech 'Speech Data Explorer' component (affected versions 0.0 through 2.9, all platforms) lets an attacker who supplies crafted malicious data achieve code execution in the context of the user running the tool, with follow-on privilege escalation, information disclosure, and data tampering. The flaw requires no authentication on the target (PR:N) but is local in reach (AV:L) and depends on user interaction (UI:R) - a user must actively open or load the attacker-crafted dataset, manifest, or serialized artifact inside the Explorer, which confines exposure to developers and researchers running NeMo rather than production network services. No public exploit code was identified at time of analysis, there is no confirmed active exploitation (the CVE is not in CISA KEV), and the CVSS 3.1 base score is 7.8 (CWE-502).
Code execution, data tampering, denial of service, and information disclosure in NVIDIA NeMo Speech (versions 0.0 through 2.9) stem from unsafe deserialization of parameters injected through a maliciously crafted model_config.yaml handled by the dataset-loading workflow. The issue is locally scoped: the assessed vector is CVSS:3.1/AV:L/AC:L/PR:L/UI:R with a base score of 7.8, meaning an attacker with low privileges must place the crafted config where the victim or an automated pipeline will parse it - for example inside a shared or downloaded model bundle or dataset - and exploitation depends on that untrusted NeMo config being loaded. There is no remote network-service exposure, and no public exploit identified at time of analysis; the CVE is not listed in CISA KEV.
Opening a maliciously crafted file in Adobe Substance 3D Modeler can trigger a stack-based buffer overflow that lets an attacker execute arbitrary code in the context of the current user. The flaw is client-side and local in nature - there is no network-facing exposure - so exploitation depends on social engineering to convince a victim to open a poisoned file, and any code that runs inherits only that user's privileges (no privilege escalation). The CVSS 3.1 base score of 7.8 (AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H) matches that assessment: high confidentiality and integrity impact, but gated behind the requirement that the user opens the file; no public exploit code or confirmed active exploitation was identified at time of analysis.
Opening a crafted file in Adobe Bridge can trigger a heap-based buffer overflow that results in arbitrary code execution with the privileges of the user running Bridge (CVE-2026-75649, CVSS 7.8). The attack is local-host only and depends on user interaction - a victim must open an attacker-supplied malicious file - so it cannot be triggered remotely or automatically, and code executes in the current user's context with no privilege escalation (AV:L/PR:N/UI:R/S:U). At time of analysis there is no confirmed active exploitation (not in CISA KEV) and no public exploit code identified, but because the advisory is vendor-published and the flaw is a memory-corruption RCE class, it should be treated as a genuine but conditional risk requiring social engineering to deliver.
Stack-based buffer overflow in Adobe Bridge lets a crafted file achieve arbitrary code execution with the privileges of the user who opens it (CVSS 7.8, CWE-121). The attack is local and requires user interaction - an attacker must deliver a malicious file and convince the victim to open it in Bridge (AV:L/AC:L/PR:N/UI:R/S:U), after which code runs in the current user's context without privilege escalation. Impact is limited by the strong social-engineering and delivery prerequisites, and no public exploit identified at time of analysis; risk is concentrated in users who routinely open files from untrusted or external sources.
Heap-based buffer overflow in Adobe Bridge lets an attacker execute arbitrary code in the context of the current user when a victim opens a specially crafted file. The flaw carries a CVSS 3.1 base score of 7.8 (AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H), meaning the attacker needs no prior privileges but must rely on social engineering to get the target to open a malicious file locally; successful exploitation yields full confidentiality, integrity, and availability impact within the user's own security context, with no privilege escalation to system level. There is no public exploit identified at time of analysis and the issue is not confirmed actively exploited (CISA KEV), so this is a genuine but non-urgent priority that should be handled through routine vendor patching; Adobe has published an advisory (APSB26-148) with fixed builds.
Arbitrary code execution in Adobe Bridge occurs when a victim opens a specially crafted file, allowing attacker-controlled code to run with the privileges of the current user. The issue is scored 7.8 (HIGH) and is a client-side, file-format-triggered flaw with a local attack vector and mandatory user interaction (CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H); the attacker does not authenticate to the product, but delivery depends entirely on social engineering the victim into opening the malicious file. Only unpatched Adobe Bridge versions addressed by vendor advisory APSB26-148 are affected, and no public exploit identified at time of analysis.
Out-of-bounds write in Adobe Substance3D Modeler allows arbitrary code execution in the context of the current user when a victim opens a malicious file. This local attack (CVSS 3.1 vector AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H, base 7.8) requires user interaction and is not known to be actively exploited; no public exploit code has been identified at time of analysis. The flaw affects the application's file parsing logic and could lead to full compromise of confidentiality, integrity, and availability for the current user.
An out-of-bounds write in Adobe Substance3D Modeler lets an attacker execute arbitrary code in the context of the current user when a victim opens a specially crafted file in the application. The flaw is triggered during parsing of a malicious asset, so exploitation is local and requires user interaction: there is no network-reachable or drive-by path, and no public exploit code has been identified at time of analysis. With a CVSS base of 7.8 (AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H) and no CISA KEV listing, this is a genuine but lower-priority-than-the-headline-number memory-corruption bug whose real-world impact is bounded by the privileges of the user who opens the file.
Opening a maliciously crafted 3D model or project asset in Adobe Substance 3D Modeler can trigger an out-of-bounds write during file parsing, allowing an attacker to corrupt memory and execute arbitrary code under the privileges of the user running the application. The flaw carries a CVSS 3.1 base score of 7.8 (AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H): no privileges are required (PR:N), but exploitation is gated by mandatory user interaction - the victim must open the hostile file - and the impact is confined to the current user's session rather than escalating to SYSTEM or root. There is no evidence of active exploitation (not listed in CISA KEV) and no public exploit code identified at time of analysis, so risk is limited to environments where users routinely open untrusted 3D asset files; Adobe has published vendor guidance in advisory APSB26-155.
Out-of-bounds write (CWE-787) in Adobe Bridge allows an attacker to achieve arbitrary code execution in the context of the current user when a victim opens a specially crafted file. Adobe rates the flaw 7.8 (HIGH) under CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H - a local, client-side vector that requires user interaction rather than any network-facing or service exposure, with PR:N simply meaning no application authentication is needed while social engineering remains a prerequisite. No public exploit code or confirmed active exploitation was identified from the available intelligence at time of analysis, and the affected version range is unspecified in the source data, so any deployed Adobe Bridge build should be assumed potentially vulnerable until the vendor advisory is reconciled against the installed release.
Out-of-bounds write in Adobe Bridge can be triggered when a user opens or previews a maliciously crafted file, allowing memory corruption that may lead to arbitrary code execution in the context of the current user. Only systems with Bridge installed where users routinely open untrusted asset files are exposed; the vector is local (AV:L) with mandatory user interaction (UI:R), so there is no unattended or remotely reachable attack path and no privilege escalation beyond the logged-in user (S:U). Adobe self-reported the flaw in advisory APSB26-148 and rates it CVSS 3.1 7.8 HIGH; no public exploit code was identified at time of analysis and there is no KEV status confirming active exploitation.
Code injection in NVIDIA NeMo Speech versions 0.0 through 2.9 allows local attackers with low privileges to execute arbitrary code, disclose sensitive information, and tamper with data by supplying crafted input to the toolkit. The vulnerability requires local access (AV:L) and low privileges (PR:L) but no user interaction; it has a CVSS 3.1 base score of 7.8 (High) and is classified as CWE-77 (command injection). No public exploit code has been identified at time of analysis, and the vendor advisory at GitHub NVIDIA product-security 2026/5885 should be consulted for patch details.
Local cross-user code execution with privilege-escalation potential affects the Windows builds of GNU wget distributed by eternallybored.org (releases 0 through 1.21.4 inclusive). The flaw is a hardcoded configuration location, C:\msys64, that unprivileged users can write to: a low-privileged local attacker who plants a wgetrc there can abuse the use_askpass directive so that arbitrary code is executed in the security context of a higher-privileged user who later runs the affected wget build. The attack is not remotely reachable and does not trigger on its own, which is captured by the UI:R component of the CVSS 3.1 vector (AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H, base 7.8), and no public exploit code was identified at time of analysis.