Information disclosure in PMWeb v7.x, v8.x, and v2025.x allows a remote, unauthenticated attacker to extract sensitive data by manipulating an unspecified function inside the encryptionhelper.dll library. The vulnerability was disclosed to the vendor, who did not respond, so no vendor-released patch has been identified at the time of analysis and no public exploit code was identified either. The confidentiality impact is limited (VC:L) with no integrity or availability effect, making this a low-priority issue in practice - real-world exposure depends on whether a given PMWeb instance is reachable from untrusted networks, since the product is normally deployed behind corporate access controls.
The F&F Filipowski mH-DEVELOPER smart home module ships production firmware built on end-of-life Debian 8 and Node.js 17.0.1, both of which are past their support lifecycles and carry publicly documented, permanently unpatched vulnerabilities. Rather than containing one discrete exploitable flaw, the product is exposed because it bundles obsolete, unsupported third-party components (CWE-1104); an attacker who can reach and correctly trigger one of the known component CVEs in this firmware could achieve arbitrary code execution, data disclosure, or denial of service, and the network-reachable services require no authentication (CVSS vector signals PR:N). No public exploit identified at time of analysis, and no KEV listing, so this is best treated as a software-composition hygiene finding whose real-world risk is contingent on how exposed the device's services are; F&F addressed the issue in firmware 3.0.30.
Unauthenticated remote attackers can call the cleanCache handler in NotionNext's pages/api/cache.js to force cache invalidation, degrading availability for sites built on the framework. Every release from 4.1.0 through 4.10.10 is affected: versions 4.1.0-4.9.5.2 perform no HTTP method or authentication check on the endpoint at all, while 4.9.5.7-4.10.10 add a guard that is only enforced when the CACHE_REVALIDATION_TOKEN environment variable is explicitly set, so default deployments remain exposed. Impact is confined to availability (assessed CVSS 3.1 AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L) with no data read or write and no code execution; no public exploit code or active exploitation has been identified at time of analysis, and the vendor was contacted but did not respond, so no fix has been confirmed.
Dell BOSS-N1 controllers running firmware prior to 2.2.13.2038 expose an unprotected on-chip debug and test interface on the SMCU, allowing anyone with physical access to bypass authentication and read or write protected controller memory, yielding high confidentiality, integrity, and availability impact (CVSS 6.8, AV:P/AC:L/PR:N/UI:N). The attack surface is strictly local: exploitation requires hands-on connection to the debug interface of the SMCU on a 17th-generation PowerEdge server, so remote or network attack is impossible and realistic exposure is limited to insiders, hardware in transit or during servicing, decommissioned or resold units, and evil-maid/supply-chain scenarios. No public exploit code or confirmed active exploitation has been identified at time of analysis, and the physical-access prerequisite keeps practical priority low despite the Medium base score.
ServiceAccount token exfiltration in the StreamsHub Console for Apache Kafka lets a user who already holds RBAC permission to create or modify the Console custom resource point the console-api AdminClient at an attacker-controlled broker. Tenant-supplied Kafka client properties are copied into the AdminClient configuration without filtering security-sensitive keys, so setting config.providers and bootstrap.servers causes the console-api ServiceAccount token to be read and leaked off-cluster. The issue is rated CVSS 6.5 (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N, CWE-470) and is a moderate-severity, privilege-dependent information disclosure: it requires an authenticated, semi-trusted role with write access to Console CRs (PR:L), so clusters with tight RBAC on those resources are substantially less exposed, and unauthenticated external actors cannot trigger it. No public exploit code has been identified at time of analysis and the flaw is not confirmed actively exploited (not in CISA KEV); no EPSS score was provided.
Relative path traversal (CWE-23) in SUSE Rancher Fleet allows an authenticated attacker with either Git push access to a repository referenced by a GitRepo resource or Kubernetes RBAC permission to create or modify a GitRepo (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N) to read files from the filesystem of the bundle-processing environment and have their contents emitted into the generated Bundle resource. The most valuable target is Helm registry credential material that Fleet makes available to the bundle-processing job when per-path Helm credentials are configured, letting the attacker exfiltrate secrets they have no Kubernetes RBAC permission to read. All Fleet branches below 0.16.2, 0.15.7, 0.14.11, 0.13.16 and 0.12.20 are affected; there is no public exploit identified at time of analysis and no evidence of confirmed active exploitation, so this is a moderate privilege-boundary-crossing disclosure rather than an emergency.
Arbitrary file write via path traversal in the kubectl client on Windows allows a malicious tar binary inside an untrusted container to place files at attacker-chosen locations on the operator's workstation, limited only by the permissions of the local user running kubectl. Exploitation requires victim interaction: the user must run `kubectl cp` to copy files FROM an attacker-controlled container, and only affects kubectl clients on Windows (v1.34.0-1.34.11, v1.35.0-1.35.8, v1.36.0-1.36.4) - Linux and macOS clients are explicitly unaffected. The Kubernetes Security Response Committee rated this Medium at CVSS 6.5, our independent assessment concurs at CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:N/I:H/A:N (unauthenticated, user-interaction required, high integrity impact only), and no public exploit code was identified at time of analysis.
Remote, unauthenticated attackers who control an X server can crash X clients linked against libX11 before 1.8.14 by returning crafted XIM (X Input Method) attribute data, because the parser performs an out-of-bounds read (CWE-125). This is an availability-only denial-of-service (CVSS 6.5, AV:N/AC:L/PR:N/UI:R/S:U/C:N/I:N/A:H) limited by an inverted trust relationship: the victim's client must attach to the attacker's display, so exploitation requires the user to connect to a malicious, compromised, or spoofed X server. No public exploit code has been identified at time of analysis, and the issue is not listed in CISA KEV.
SUSE Rancher Fleet versions 0.16 before 0.16.2, 0.15 before 0.15.7, 0.14 before 0.14.11, and 0.13 before 0.13.16 allow an attacker with low-privilege bundle deployment rights to modify labels and annotations on Kubernetes namespaces that the bundle's pinned identity is not authorized to change. The flaw arises from inconsistent authorization checks when a bundle requests namespace metadata updates via the namespaceLabels and namespaceAnnotations options. No public exploit code has been identified at time of analysis, and the CVSS 3.1 score of 6.5 reflects integrity-only impact (C:N/I:H/A:N) requiring existing low privileges on the network.
Authenticated path traversal in Bimser eBA Plus Document and Workflow Management System (versions 6.7.141 up to but not including 10.0.11) allows a remote attacker holding a valid low-privilege account to read files outside the intended restricted directory, resulting in high confidentiality impact with no integrity or availability effect (CVSS 3.1 base 6.5, AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N). The vulnerability is a CWE-22 improper limitation of a pathname to a restricted directory, reported by TR-CERT, and exploitation requires an authenticated session - anonymous access is not indicated by the assessed vector. No public exploit code has been identified at time of analysis, and the flaw is not listed in CISA KEV, so this is a moderate-severity, credentialed information-disclosure issue rather than a top-priority emergency; because the vector is PR:L, only confirmed as an authenticated attack path. Deployments on 10.0.11 or later are not affected.
Missing authentication in the mH-DEVELOPER smart home module lets any device on the same LAN query its HTTP API and WebSocket endpoints and issue raw control commands to building-automation hardware, because the authorization middleware never verifies tokens. The flaw maps to CWE-306 (Missing Authentication for Critical Function) and affects every release prior to 3.0.30, the version in which the vendor fixed it; no public exploit code has been identified at time of analysis. The vendor CVSS 4.0 vector (AV:A/AC:L/PR:N/UI:N, base 6.4) and our independent CVSS 3.1 assessment agree that exploitation is trivial but gated by network adjacency: the attacker must already be on the same local or building network, and no authentication or user interaction is required beyond that.
Trust-boundary bypass in the ip-address JavaScript library (all versions prior to 10.5.1) lets a remote, unauthenticated attacker who can supply or influence the IPv6 address being validated slip an address within the full fe80::/10 link-local range past Address6.isLinkLocal(), which only matched the narrower fe80::/64. Applications that use isLinkLocal() as an SSRF or on-link filter - the sole prerequisite for exploitation, reflected as AT:P in the CVSS 4.0 vector and AC:H in the independent assessment - may then be steered to reach a link-local host outside the intended trust boundary, with impact limited to confidentiality and reachability (C:L, S:C) and no integrity or availability effect. No public exploit code identified at time of analysis; the vendor fix is version 10.5.1.
HTTP/1 response smuggling in the Elixir Mint client (0.1.0 through 1.10.x) allows a malicious or attacker-influenced origin server to desynchronize Mint from an upstream intermediary on a pooled keep-alive connection, so that bytes following a crafted response are parsed as the answer to a subsequent request, poisoning that request's response. The root cause is that Mint's message_body/1 selects chunked framing whenever 'chunked' is the FIRST coding in a Transfer-Encoding header (e.g. 'Transfer-Encoding: chunked, gzip') instead of requiring it to be the final coding, and additionally keeps the connection open after an HTTP/1.0 response - final or 1xx - that carries Transfer-Encoding with 'Connection: keep-alive', both contrary to RFC 9112. Exploitation is a reverse-trust scenario: the attacker must control or inject server response framing and the client must reuse an HTTP/1 connection; HTTP/2 and non-pooled connections are unaffected, and no public exploit code or active exploitation was identified at time of analysis. The issue is fixed in Mint 1.11.0.
Memory exhaustion in the Elixir Mint HTTP/2 client (mint 0.1.0 through 1.10.x) lets a malicious or man-in-the-middle HTTP/2 server pin roughly 16 MiB of client buffer memory per open connection by declaring an oversized frame and withholding its final byte. The weakness is an unbounded resource allocation issue (CWE-770) in Mint.HTTP2.Frame.decode_next/2, which only checks a frame against the client's 16,384-byte max_frame_size after the entire declared payload has arrived, causing the connection buffer to keep accumulating bytes up to the 24-bit frame-length ceiling of 16,777,215 bytes. Impact is availability-only and low severity per the independent assessment: there is no amplification because the attacker must transmit every byte the client buffers, sustained denial of service requires many concurrently held connections plus matching attacker resources, and no public exploit identified at time of analysis, with no CISA KEV listing and no EPSS signal supplied. Attacker control of the server side is required (unauthenticated from the client's perspective, reflected as AT:P), and vendor patches are available in mint 1.11.0.
Privilege escalation in Apache Karaf versions before 4.4.12 allows any authenticated user of the shell or SSH console - including one holding only the low-privilege 'viewer' role - to abuse the config:install command and write attacker-controlled content into the sensitive ${karaf.etc} directory, escalating to full administrative control of the runtime. The flaw exists because config:install has no entry in the shipped etc/org.apache.karaf.command.acl.config.cfg ACL while unmatched commands fail open, and because karaf.secured.command.compulsory.roles ships commented out in etc/system.properties; with the -o/--override flag an existing file such as users.properties, keys.properties or host.key can be overwritten, and Felix FileInstall (felix.fileinstall.dir=${karaf.etc}) hot-reloads any dropped .cfg file without a restart. The vendor rates this 'moderate'; no EPSS score, no CISA KEV entry and no public POC were provided, and no public exploit has been identified at time of analysis. Authentication is required and the vendor-fixed version is Apache Karaf 4.4.12.
Canonical LXD before 5.0.10, 5.21.8, and 6.10 lets a project-restricted client read its own private images from other projects through local fingerprint reuse during image or instance import requests. The flaw is a missing-authorization condition (CWE-862) graded CVSS 6.3 (AV:N/AC:H/PR:L/UI:N/S:C/C:H/I:N/A:N), meaning an attacker needs a valid but low-privileged credential and a multi-tenant deployment where a private image is already cached locally by another project; impact is limited to confidentiality of other projects' image content, with no code execution, write, or denial-of-service effect. No public exploit code or confirmed active exploitation was identified at time of analysis, and vendor patches are available. This is a genuine but moderate-severity project-isolation weakness rather than a critical one, and single-project or non-shared LXD hosts are not exposed.
Man-in-the-middle attacks against F&F Filipowski mH-DEVELOPER smart home modules running firmware before 3.0.30 are possible because every unit ships with the same hard-coded SSH host keys and performs no per-device key generation, so an attacker who extracts those keys from a firmware image can stand up a rogue SSH server that clients accept without any warning. Exploitation is unauthenticated (PR:N) but requires two prerequisites in combination: possession of the static keys and an on-path position between an SSH client and the device (reflected in the CVSS 4.0 AT:P requirement, and in our assessed CVSS:3.1 AC:H), which keeps this a moderate, on-path-dependent flaw rather than a mass-exploitable one; no public exploit code or confirmed active exploitation was identified at time of analysis. Impact is limited to credential and data interception (VC:L/VI:L) with no availability effect, and devices updated to 3.0.30 are not vulnerable.
Agentis versions 4.44 up to (but not including) 4.6 are vulnerable to a reflected cross-site scripting (XSS) flaw (CWE-79) that allows an attacker to inject script into HTML attribute contexts within generated web pages. Exploitation requires a victim to load attacker-controlled input, typically via a crafted link, and no authentication is required for the attacker; the impact is limited to the victim's browser session with low confidentiality and integrity impact. No public exploit code has been identified at time of analysis, and the vendor CVSS 3.1 base score is 6.1.
Cleartext transmission of credentials, authentication tokens, and device commands in the mH-DEVELOPER smart home module (F&F Filipowski) exposes every web interface and API session to interception by anyone sharing the same local network segment. Every version prior to 3.0.30 is affected, and because no TLS is used, a passive sniffer or an attacker who establishes an interception position (for example via ARP spoofing or a rogue access point) can harvest passwords and session tokens and then hijack sessions to issue device commands. Exploitation requires only an adjacent-network position and no authentication on the module itself; there is no confirmed active exploitation (not listed in CISA KEV) and no public exploit identified at time of analysis, but the confidentiality impact is high and the flaw is trivially exploitable wherever untrusted clients, guest Wi-Fi, or shared VLANs can reach the device. Vendor-released patch: version 3.0.30.
Denial of service in F&F Filipowski mH-DEVELOPER smart home modules running versions prior to 3.0.30, where the Node.js-based fh-node service accepts JSON and URL-encoded bodies up to 250 MB through the Express bodyParser, allowing a LAN-adjacent attacker to exhaust device RAM and crash the process with out-of-memory conditions. The vendor's CVSS 4.0 vector rates this 5.9 (medium, AV:A/AC:L/AT:P/PR:L) and is availability-only with no confidentiality or integrity impact; successfully triggering the crash is probabilistic because it depends on the device's instantaneous memory usage, which the attacker does not fully control, hence the AT:P rating. Although the vector nominally requires privileges (PR:L), the accompanying assessment notes that sibling flaw CVE-2026-82930 allows all endpoints to be queried without authentication, so any LAN user can reach the vulnerable body-parsing path in practice. No public exploit code was identified at time of analysis and the flaw is not in CISA KEV; exploitation is bounded to adjacent-network reach rather than the internet. Vendor-released patch: 3.0.30.
Arbitrary file and directory-tree writes on an LXD client host can be achieved through the recursive file pull feature of the LXD CLI in Canonical LXD 4.0.2 through 6.9 on Linux. An attacker who already holds root inside a guest VM (CVSS PR:L) substitutes a modified lxd-agent that returns inconsistent SFTP directory listings and Lstat results, then waits for the host operator to invoke a recursive pull against that VM (UI:R); the difficulty of forging those inconsistent responses is captured by AC:H, and the write lands on the host with the operator's privileges (S:C, I:H, C:N, A:N). This is CWE-59 improper link resolution, no public exploit code was identified at time of analysis, and vendor patches are available in 4.0.14, 5.0.10, and 5.21.8.
Sandboxed Flatpak applications can deliver signals to a process group that also contains a parent process outside the app's sandbox, letting a malicious or compromised app terminate the desktop environment or user session (denial of service) on Flatpak versions before 1.18.4. Exploitation is local and gated on the victim installing and executing attacker-controlled sandboxed code; the assessed CVSS v3.1 vector (AV:L/AC:L/PR:L/UI:N/S:C/C:N/I:N/A:H) records a scope change with high availability impact and no confidentiality or integrity impact. No CISA KEV listing, EPSS score, or public exploit code was identified at time of analysis, and the issue is fixed in Flatpak 1.18.4, released alongside sibling fixes CVE-2026-97023 through CVE-2026-97027.
Path traversal in PMWeb 7.x, 8.x, and 2025.x lets unauthenticated remote attackers read files from the server filesystem by supplying traversal sequences in the FullFileName or FileName argument of downloader.aspx. The issue is rated CVSS v4.0 6.9 and independently assessed as CVSS v3.1 6.9 (AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N), i.e. a network-reachable, no-authentication, read-only file disclosure with high confidentiality impact and no integrity or availability impact. Publicly available exploit code exists, but there is no confirmation of active exploitation (not in CISA KEV), and the vendor did not respond to the disclosure; practical risk is tempered by PMWeb's niche deployment in construction and capital-program organizations and the frequent internal or access-restricted placement of those instances.
Dell Live Optics Collector versions prior to 27.2.13.310 ship with a hard-coded password (CWE-259) that a locally present, low-privileged attacker can recover and abuse to expose sensitive information. Exploitation is constrained: it requires local access to a host where the Collector is installed, an existing low-privileged account (PR:L), and user interaction (UI:R), yielding high confidentiality impact but only limited integrity impact and no availability impact (CVSS 3.1 base 5.6). There is no confirmed active exploitation (not in CISA KEV) and no public exploit code identified at time of analysis; Dell has published fix version 27.2.13.310 in advisory DSA-2026-339.
Path traversal in ABB Protection and Control IED Manager (PCM600) through version 2.14 allows a local, low-privileged attacker to overwrite or corrupt files on an engineering workstation by supplying a crafted project, configuration, or import file with manipulated pathnames. Exploitation requires the attacker to already have local access with at least low-privilege credentials (PR:L) and depends on user interaction (UI:R) - an engineer must open or import the malicious file - and impact is limited to integrity (C:N/I:H/A:N) with no confidentiality or availability effect, making this a moderate, contained risk rather than a critical priority. No public exploit code has been identified at time of analysis, and there is no confirmed active exploitation (CISA KEV) or EPSS score provided for this CVE.
A stack-based buffer overflow in the Cesanta Mongoose MQTT broker tutorial (tutorials/mqtt/mqtt-server/main.c) allows remote, unauthenticated attackers to crash the broker process by sending a manipulated MQTT message to the example's `fn` handler. Only deployments that build and expose this specific tutorial server at version 7.21 or earlier are exposed - the flaw lives in example code shipped with the repository rather than the core embedded networking library, so applications that embed Mongoose without compiling this sample are not affected. Publicly available exploit code exists, and the assessed impact is limited to denial of service, with no code execution or data disclosure indicated.
Crashing of X11 client applications can be triggered by a malicious X server exploiting an out-of-bounds read in libX11's X Input Method (XIM) trigger-key registration parser in versions before 1.8.14. This is a medium-severity denial-of-service issue (CVSS 5.5) that requires the victim to connect to an attacker-controlled X server, with user interaction (UI:R) and no authentication required (PR:N), but impact is limited to availability (client crash). No public exploit identified at time of analysis, and it is not confirmed actively exploited (CISA KEV).
An unsigned integer underflow in libXpm's XPM write/save path (libXpm before 3.5.19) causes unbounded CPU consumption and memory exhaustion, resulting in denial of service. The flaw is reachable only when an application linked against the library is made to write or convert attacker-influenced data to the XPM format, so exploitation is local (AV:L) and requires the victim to actively process the crafted input (UI:R); no privileges are required (PR:N) and there is no path to data disclosure or code execution. No public exploit has been identified at time of analysis, and no confirmed active exploitation (CISA KEV) has been reported; the upstream fix is present in a libXpm merge-request commit, with libXpm 3.5.19 indicated as the corrected version. Given the local, interaction-dependent trigger and DoS-only impact (CVSS 5.5), this is a genuine but low-priority issue.
Local denial of service in Wind River VxWorks 7 (all releases prior to 26.09) lets a low-privileged local task exhaust kernel memory by invoking process-management system calls whose specific arguments cause the subsystem to skip freeing allocated kernel memory when the calling application terminates. Repeated invocation accumulates the leak until the kernel memory pool is depleted, producing a high availability impact (CVSS 5.5, AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H). No public exploit code has been identified at time of analysis and the issue is not confirmed as actively exploited; Wind River has fixed it in VxWorks 7 version 26.09.
Memory pool exhaustion in Wind River VxWorks 7 releases 24.03 through 26.03 can terminate system operations when specific service routines are invoked while the device is in a non-default configuration state, because allocated memory pools are not released before the failure propagates. The weakness is local and availability-focused (CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:H per independent assessment): an attacker needs low-privileged local execution on the RTOS and the affected services must be reachable, and no user interaction or network path is required. Default installations are not affected, the configuration prerequisite plus the local-access requirement substantially narrow the realistic attacker population, and no public exploit identified at time of analysis; Wind River has fixed the issue in VxWorks 7 26.09. Overall this is a moderate-severity, low-real-world-priority issue rather than a broadly exploitable one.
Local resource exhaustion in Wind River VxWorks 7 (versions prior to 26.09) allows a low-privileged local application to leak kernel memory and system file descriptors by invoking system calls with specific argument values, causing the IPNET subsystem to fail to release those resources before the calling application terminates. Per the assessed vector (CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H, base 5.5), an attacker must already be executing code on the device, so this is an availability-only denial-of-service rather than a remote or unauthenticated attack, and there is no confidentiality or integrity impact. There is no public exploit identified at time of analysis and no confirmed active exploitation (CISA KEV); Wind River has released a fixed version, 26.09, so the primary remediation is a firmware upgrade.
In Trusted Domain Project OpenDMARC 1.4.2 and earlier, unauthenticated remote attackers can trigger improper handling of an exceptional condition in the strcasecmp-based subdomain-walk logic of libopendmarc/opendmarc_policy.c, degrading DMARC policy evaluation. A publicly available exploit (including a technical writeup) exists, and the vendor was contacted but did not respond, so no vendor-released patch is identified at time of analysis. Impact is limited to low integrity and low availability effects within DMARC processing - no confidentiality exposure and no code execution - meaning the practical risk is spoofing/phishing that evades DMARC enforcement at sites that rely on OpenDMARC as their anti-spoofing milter.
An off-by-one error in OpenDKIM's quoted-printable decoder (dkim_qp_decode in util.c) affects all releases up to and including 2.11.0 and is reachable by remote, unauthenticated attackers who can deliver mail to a host running OpenDKIM as a verifying milter. The flaw is triggered while decoding attacker-controlled message bodies, producing low-magnitude confidentiality, integrity and availability impact rather than reliable code execution, consistent with the assessed CVSS 3.1 vector AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L. Publicly available exploit code exists, and the vendor was contacted before disclosure but did not respond, so no vendor-released patch has been identified at time of analysis.
Rancher Fleet 0.16.0 and 0.16.1 accept requests to the gitjob webhook receiver without verification whenever no webhook secret is configured, allowing any party with network reach to that endpoint to alter the spec.pollingInterval field of matching GitRepo resources in any namespace - including namespaces where the caller holds no Kubernetes credentials or RBAC. The assessed vector is unauthenticated remote (AV:N/PR:N), and impact is deliberately narrow: only a single configuration field can be changed, with no arbitrary spec edits, secret exposure, or workload execution and no confidentiality loss. Only Fleet 0.16.0/0.16.1 are affected, Fleet 0.16.2 and all pre-0.16 releases are not, and no public exploit code has been identified at time of analysis.
PyJWT versions prior to 2.14.0 accept signature segments that are not canonical Base64URL, so a token whose compact JWS signature has non-alphabet characters appended still decodes to the same signature bytes and passes verification. The practical impact is limited: an attacker who already holds a valid-but-revoked token can craft an equivalent serialized string that bypasses revocation or blocklisting logic which matches on the raw token string rather than a claim such as jti. This is a low-severity defense-in-depth flaw (assessed CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:L/A:N) that requires the application to compare raw serialized tokens for revocation. No public exploit identified at time of analysis.
Local code execution in Google fuse-archive versions prior to 1.24 lets an attacker hijack the external decompression helpers that libarchive launches when an archive is mounted, by prepending a directory to the process PATH or by planting a malicious binary in an attacker-writable directory that already appears on PATH. The flaw is an untrusted search path issue (CWE-426) rated 5.4 under CVSS 4.0 (AV:L/AC:L/AT:P/PR:L/UI:A/VC:H/VI:H/VA:H); exploitation requires local access as an authenticated user (PR:L), that fuse-archive run without the '-o noexternal' option, and that a victim mount an archive whose format triggers an external filter such as base64 or gpg. No public exploit code or CISA KEV entry was identified at time of analysis; version 1.22 delivered only a partial mitigation and version 1.24 fully resolves the issue by restricting external helper lookups to a sanitized, fixed PATH.
Unbounded buffering in the Server-Sent Events parser of the AG-UI Community SDK (Kotlin/Java) lets a malicious or man-in-the-middle-positioned agent endpoint drive a client-side memory exhaustion condition, degrading or killing the consuming process. All ag-ui-protocol ag-ui builds up to 2026-09-23 are flagged by CPE, specifically the Kotlin Community SDK SseParser.kt and its Java counterpart SseEventParser, which had no per-line or per-event size caps before the fix. Impact is availability-only (no confidentiality or integrity effect), an upstream fix exists as PR #2657 but was still awaiting acceptance, and no public exploit was identified at time of analysis.
Stored cross-site scripting (XSS) in eBA Plus Document and Workflow Management System (versions 6.7.141 up to but not including 10.0.11) allows an authenticated low-privilege user to persist script payloads into document, form, or workflow content that other users later render in their browser, where the script executes in the victim's session context. The assessed vector confirms the practical constraints: a valid account is required (PR:L) and a separate user must open the tainted record for the payload to fire (UI:R), with the impact crossing a trust boundary (S:C) so the script can act as the viewing user - this is therefore not exploitable against unauthenticated or default anonymously accessible deployments. Reported through TR-CERT (TR-26-1197), the issue carries a moderate CVSS base score of 5.4, and no public exploit code or confirmed active exploitation (CISA KEV) was identified at time of analysis.
Reflected cross-site scripting in the OrdaSoft Book Library (Free) extension for Joomla, affecting versions below 6.4.6, lets an unauthenticated remote attacker inject arbitrary HTML and JavaScript that executes in a victim's browser session. The flaw lives in the publicly reachable book-detail view (site/views/view_book/tmpl/default.php), where the 'title' request parameter is echoed raw inside a double-quoted HTML attribute, so a value containing a double quote breaks out of the attribute and everything after it is rendered as markup. Exploitation requires passive user interaction - the attacker must deliver a crafted link to a victim (UI:P) - and impact scales with the victim's privileges, so a lured logged-in administrator yields a far more serious outcome than an anonymous visitor; no confirmed active exploitation (not in CISA KEV) and no public exploit code identified at time of analysis, and the first-party assessment rates this a genuine but modest-severity reflected XSS.
Reflected cross-site scripting in the Vehicle Manager (Free) extension for Joomla before 6.5.8 allows an unauthenticated attacker to inject arbitrary markup and script into a page by crafting a malicious 'title' parameter on the public vehicle-detail page (task=view). The parameter is echoed straight into a double-quoted HTML attribute with no output encoding, so a double-quote character breaks out of the attribute and permits a <script> tag to be injected; exploitation requires the victim to follow a crafted link (CVSS 4.0 vector shows PR:N with UI:P, yielding a base score of 5.3). No public exploit code or confirmed in-the-wild exploitation has been identified at time of analysis, so this is a genuine but low-severity reflected XSS whose practical impact is limited to script execution in the context of a victim user's session.
Reflected cross-site scripting in the ordasoft.com Real Estate Manager (Free) extension for Joomla before version 6.7.9 lets remote attackers execute arbitrary script in a victim's browser by luring them to a crafted URL. The flaw lives in the 'leave a review' form on the public property-detail page, where the review title is echoed from the request straight into a double-quoted HTML attribute with no escaping or filtering at all - a double-quote character in the title parameter breaks out of the attribute and permits a following <script> element to run. No authentication is required and the CVSS 4.0 base score is 5.3 (assessed as CVSS 3.1 AV:N/AC:L/PR:N/UI:R/S:C/C:L/I:L/A:N); exploitation depends on the victim loading the malicious link, so it is not self-triggering, and no public exploit code was identified at time of analysis. Impact is contained to the tricked user's browser session, with no server-side compromise; this is a genuine but low-severity issue.
Broken access control in ZoneMinder's Frames API lets an authenticated user who holds the 'Events' view permission but is explicitly denied access to one or more monitors enumerate frame and event metadata belonging to those restricted monitors. Affected are all ZoneMinder releases before 1.38.4 whenever per-monitor access restrictions are configured; the flaw is a single authenticated GET-style call to the FramesController index endpoint, requires no user interaction and no administrative rights, and is confirmed by the CVSS 3.1 vector AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:N as a low-confidentiality-impact information disclosure only. No public exploit code and no confirmed active exploitation were identified at time of analysis, and the issue is largely irrelevant on single-tenant or fully trusted deployments where every user may already view every monitor.
An out-of-bounds read in the Zephyr RTOS CoAP link-format helper match_path_uri() (subsys/net/lib/coap/coap_link_format.c) allows an unauthenticated remote attacker to drive the URI comparison loop one or more bytes past the end of a Uri-Query href= option value by sending a GET /.well-known/core?href=/<prefix> request whose value is a strict prefix of a longer registered resource path. Only devices running the Zephyr CoAP server subsystem with CONFIG_COAP_SERVER_WELL_KNOWN_CORE enabled (the default) and at least one resource declaring struct coap_core_metadata attributes reach the vulnerable code path. The over-read is bounded - in practice a single byte, limited by the compared path length - reads only uninitialized stack tail rather than receive-buffer or attacker-controlled data, cannot influence the response, and yields undefined behaviour rather than information disclosure or a matching error; no public exploit code was identified at time of analysis, and the assessed severity is Low (CVSS 3.7).
Sandbox containment failure in the deepseek-ai deepseek-harness Code Mode Sandbox (up to and including 0.1.0-rc.7) allows remote, low-privileged attackers who can submit code through the run_code function to escape the worker's intended containment. The vendor's own SAFETY.md and design notes state the worker is "containment, not a security boundary," so the flaw primarily harms deployments that mistakenly rely on it for isolation; there is no public exploit identified at time of analysis and the pre-release, niche nature of the library keeps real-world risk low-to-moderate. Impact is limited to low confidentiality, integrity, and availability effects on the harness itself, with no downstream-scope compromise.
Stored cross-site scripting (XSS) in httpdbg before 2.2.1 allows an attacker who controls or influences a URL captured by the tool to inject a javascript: scheme link that executes script in the application origin when a user clicks it (CVSS:4.0 5.3, UI:P, PR:N, so unauthenticated but requiring user interaction). Successful exploitation exposes the recorded request and response data httpdbg displays - including headers and captured tokens - because the malicious script runs in the origin of the local debugging web interface. The issue is not reported as actively exploited and no public exploit code has been identified at time of analysis, and the practical exposure is narrowed by httpdbg's use as a developer-facing, typically local debugging tool that a victim must open and interact with; vendor-released patch: 2.2.1.
Cross-domain authenticated request bypass in the connect-xcors (connect-cors) Node.js middleware lets an attacker defeat origin verification so that a victim's browser issues credentialed cross-origin requests the application would otherwise reject. The flaw is remotely reachable without attacker authentication (CVSS 4.0 PR:N) but requires the victim to be logged in to the affected application and to load attacker-controlled web content (UI:P); impact is bounded to low confidentiality and low integrity with no availability effect, giving low-to-moderate real-world risk. No public exploit code has been identified at time of analysis, and no CISA KEV entry exists for this CVE.
Memory-safety defect in Meta's Proxygen C++ HTTP library lets the HTTPTransaction::onWebTransportUniStream and HTTPTransaction::onWebTransportBidiStream callbacks return stream handles that the stream handler had already freed, after which HQSession installs those stale handles as transport read callbacks and can read through freed memory. Only servers built on affected releases (v2024.10.28.00 up to but not including v2026.09.28.00) that actually enable and terminate WebTransport over HTTP/3 are exposed; remote, unauthenticated peers able to drive that stream-setup path can trigger the fault, but the assessed impact is availability only (CVSS 5.3, A:L, no C/I loss) and reliable triggering depends on winning the free-then-reuse timing window. No public exploit code has been identified and there is no confirmed active exploitation; EPSS is very low at roughly 0.14% (3rd percentile).
Cross-site request forgery in Cotonti CMS through 1.0.0 lets an attacker who lures an authenticated administrator to a malicious page delete custom user groups and their associated permissions without any token verification. The vulnerable logic sits in system/admin/admin.users.php (around lines 136-140), where the user-group deletion action executes without the framework's anti-CSRF (xg) check, so a forged request simply rides the administrator's live session. From the assessed vector (CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:N/I:L/A:L) the attacker needs no credentials of their own but cannot succeed without an administrator's valid session and interaction, which keeps this a low-to-moderate priority issue; no public exploit code has been identified at time of analysis, there is no evidence of confirmed active exploitation, and the vendor has published fix code.
Uncaught exception handling in the A2A middleware of ag-ui-protocol ag-ui (all versions up to 2026-09-07) allows an attacker who can control the arguments of a tool call to crash or permanently hang an agent run stream. Supplying malformed JSON, or valid JSON that is not a usable object (null, a bare number, a string, or an array), reaches an unguarded JSON.parse path in the middleware's conversion logic and throws before any error-catching exists, leaving the run neither completed nor errored. The issue is network-reachable but requires low-level/authenticated privilege to inject the crafted tool-call arguments (CVSS 4.0 base 5.3, PR:L; independent assessment CVSS 3.1 AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:L), and impact is limited to availability of the affected run with no confidentiality or integrity effect. There is no public exploit identified at time of analysis; the vendor fixed the flaw in release 2026-09-08.
Incomplete cleanup of per-run state in the ag-ui TypeScript client SDK's FilterToolCallsMiddleware allows disallowed tool-call events from one agent run to escape filtering - or valid events from another run to be wrongly blocked - when a single middleware instance or subscription serves more than one run. Affected deployments are those using ag-ui (ag-ui-protocol) client SDK builds up to 2026-09-07 that rely on FilterToolCallsMiddleware with a disallowedToolCalls list as a real security control, in a pattern where runs overlap, stall, or reuse runId/toolCallId values. The issue is rated CVSS 4.0 5.3 (assessed as CVSS 3.1 AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:L) - an authenticated, low-privileged party who can start agent runs is the relevant actor; impact is bounded to filter bypass and mis-blocking rather than code execution, and no public exploit identified at time of analysis. Vendor-released patch: 2026-09-08.