TLS certificate verification is disabled by default in Socket Firewall (socketdev/socket-registry-firewall) before version 2.0.0, enabling network-positioned attackers to conduct man-in-the-middle attacks against the firewall's outbound connections to both the Socket API and upstream package registries. When the api_ssl_verify and upstream_ssl_verify keys are omitted from socket.yml, the generated configuration explicitly sets both to false, causing the OpenResty/Lua HTTP client to accept any certificate - including self-signed and untrusted certificates - without chain validation. A successful attacker can substitute malicious package content or alter the allow/block policy decisions the firewall enforces, effectively weaponizing the supply chain security control against the organizations it protects. No public exploit has been identified at time of analysis.
Authenticated Storm operator sessions are exposed to arbitrary web origin reads via three independent misconfigurations in Apache Storm Webapp 3.0.0: Logviewer CORS origin reflection combined with credentialed responses, a CORS filter initialization parameter mismatch that causes containers to silently apply permissive defaults, and unrestricted JSONP callback wrapping on all GET endpoints that bypasses the Same-Origin Policy entirely. Any malicious web page visited by an authenticated operator can silently exfiltrate cluster topology, configuration, and worker log data on the operator's behalf without requiring any additional privileges. No public exploit has been identified at time of analysis; vendor-released patch version 3.1.0 is available.
Rebalance operations in Apache Storm Nimbus 3.0.0 allow an authenticated attacker with topology rebalance rights to inject unauthorized blobstore map entries into configuration overrides, causing Supervisor nodes to fetch and locally cache blobs whose ACLs do not grant the caller access. A companion authorization-bypass in the listBlobs operation exposes the full blobstore key namespace to any caller reaching the Nimbus Thrift port without any subject check, supplying the key names that make the rebalance bypass actionable. No public exploit is identified at time of analysis; the vendor-confirmed fix is version 3.1.0 of org.apache.storm:storm-server.
Local privilege escalation in Apache Storm Worker Launcher 3.0.0 allows an authenticated Storm topology tenant to gain root on supervisor nodes by exploiting a TOCTOU race in the setuid-root worker-launcher binary: the binary chowns the entire worker directory to the untrusted topology user before re-reading the supervisor-written command file, creating a window in which the tenant can substitute attacker-controlled content. For Docker workers, the injected command executes with real uid 0 through an insufficiently constrained argument sanitizer that passes -v, --device, --cap-add, --security-opt, --user, --net, and positional arguments verbatim; for OCI workers, arbitrary host paths can be bind-mounted read-write and execution can be forced as another tenant's uid. No public exploit has been identified at time of analysis; a vendor-released patch is available in version 3.1.0.
Local privilege escalation to root in Apache Storm's setuid-root `worker-launcher` binary (storm-core 3.0.0) allows a tenant user with code execution on a supervisor node to redirect root-owned `lchown` and `chmod` operations to arbitrary host files via a symlink race. The binary uses FTS to walk worker directories and then re-resolves pathnames at syscall time, creating a TOCTOU window exploitable by any topology user who can write to an intermediate directory component. Because worker relaunches and blob updates repeatedly re-trigger the walk, exploitation is low-cost and repeatable; this defeats the isolation boundary that `supervisor.run.worker.as.user` is designed to enforce. No public exploit is identified at time of analysis.
Improper access control in CyberPanel before 2.4.4 allows an authenticated low-privilege user to bypass alias-domain filtering logic in the child-domain management subsystem, with potential cross-account integrity impact. The flaw stems from using a Django ORM `.filter(alais=0)` call to identify non-alias child domains - a comparison that can behave differently than a Python-level equality check - combined with a missing ownership rejection branch in `checkOwnership()` that allowed certain code paths to fall through without denying access. The CVSS S:C/I:H vector indicates the impact can cross tenant boundaries on shared CyberPanel instances. No public exploit is identified at time of analysis and the issue is not listed in CISA KEV.
Unauthenticated information disclosure in the User Registration & Membership WordPress plugin (versions 5.0 through 5.2.7) exposes arbitrary users' email addresses, profile fields, roles, and membership order details to any visitor. The plugin's membership confirmation page performs no ownership verification - an attacker who supplies another user's identifier in the request receives that user's data if the site operator has configured user smart tags on the page's message template. No public exploit confirmed in CISA KEV, though a publicly available proof-of-concept is documented by WPScan.
HTTP request smuggling in HAProxy 3.3.0-3.4.4 affects deployments running an HTTP/3 (QUIC) frontend forwarding traffic to HTTP/1.1 backends over reused pooled connections with chunked transfer encoding. The H3 multiplexer in src/h3.c emits an untrusted client-declared DATA frame length verbatim as the HTTP/1.1 chunk size before payload bytes arrive; when a remote unauthenticated attacker under-delivers payload then ends the stream, HAProxy returns the desynchronized backend connection to the idle pool, enabling bypass of path-based frontend deny rules and exposure of concurrent clients' Authorization headers as attacker-visible request body. No public exploit code has been identified, though the attack is probabilistic and freely retryable.
Null pointer dereference in Zephyr RTOS's MQTT-SN client subsystem allows a malicious or on-path network actor to crash any Zephyr device running an active MQTT-SN session by suppressing gateway keepalive responses. The defect is a missing assignment from the SYS_SLIST_PEEK_HEAD_CONTAINER macro in process_ping(): gw is never populated, remains NULL, and is subsequently dereferenced and passed to the memory slab allocator. The outcome is a kernel panic (with CONFIG_MEM_SLAB_POINTER_VALIDATE) or silent slab free-list corruption on targets where address 0 is writable. No public exploit has been identified at time of analysis, and no KEV listing exists; however, the trivial trigger condition and absence of MQTT-SN authentication make this straightforward for any party controlling or impersonating the client's gateway.
Insecure direct object reference in Rhymix CMS before 2.1.31 allows authenticated users to access and potentially delete uploaded files belonging to other documents by supplying arbitrary file serial numbers in document extra variable fields. The root cause is missing ownership validation in the extravar module: file-type extra variables accepted any file SRL from user input and rendered or deleted the referenced file without confirming it belonged to the current document. No public exploit code is identified; vendor patch is available in version 2.1.31, tracked as RVE-2026-1.
SSH key injection in Froxlor before 2.3.12 allows authenticated customers to write arbitrary lines into server-side authorized_keys files by embedding newline characters in SSH public key submissions to the SshKeys::add() API endpoint. An attacker with a valid customer account can plant SSH key entries bearing OpenSSH option directives (e.g., command=, from=, restrict) that persist beyond explicit key-deletion and access-revocation actions performed by administrators. No public exploit has been identified at time of analysis, but the vendor advisory confirms the flaw and the fix version.
Out-of-bounds read in PostGIS address_standardizer through 3.7.0 allows authenticated database users to crash the entire PostgreSQL cluster by supplying rule rows with out-of-range Weight values. The Weight field is consumed directly as an array index into the load_value array inside rules_add_rule() (src/gamma.c:338) without any bounds validation, so a Weight of -1 or >= NUMBER_OF_WEIGHTS triggers an OOB read that terminates the PostgreSQL backend process and drops all active cluster sessions. No public exploit code has been identified at time of analysis; patch commits are available upstream.
Path traversal in node-restify's serveStatic plugin exposes arbitrary server-side files to unauthenticated remote attackers on all versions up to and including 12.0.0. The serveStatic function in /lib/plugins/static.js fails to adequately sanitize user-controlled path segments, allowing traversal sequences to escape the intended static file root and return files from elsewhere on the filesystem. No vendor patch is available - the maintainer did not respond to responsible disclosure - leaving affected deployments without an official fix.
Unauthenticated file upload in Anil-matcha Open-Generative-AI (versions up to 1.0.11/2.0.0) allows any remote attacker to upload arbitrary files - including HTML, SVG, and executable types - to the application's S3 backend via the /api/upload-binary proxy endpoint. The endpoint accepted requests without any API key check and imposed no restrictions on file type or content, as confirmed by the patch commit f013270957f. No active exploitation or public proof-of-concept has been identified at time of analysis; EPSS data was not provided.
Unsafe PHP deserialization in PHPGurukul Small CRM 4.0 exposes the login endpoint to object injection attacks via the uncontrolled `geopluginURL` parameter passed directly to `unserialize()` in `/crm/login.php`. A remote, unauthenticated attacker who can supply a crafted serialized payload to the login success handler may trigger arbitrary object instantiation, with actual impact constrained by available PHP gadget chains in the application. A public proof-of-concept exploit exists, though the vendor-reported CVSS 4.0 score of 2.9 and AC:H rating indicate exploitation is difficult and real-world impact is assessed as limited integrity degradation rather than full compromise.
Unauthenticated access to the AWS API Gateway endpoint in simalexan/api-lambda-send-email-ses allows any remote attacker to invoke the SES.sendEmail Lambda function without credentials, abusing the deployer's AWS Simple Email Service account to send arbitrary emails. Affected deployments are those built from any commit up to bda6869aa81371d1e872242e74fe7d953edb818d, with no patched release available as the project uses rolling delivery and the maintainer has not responded to the disclosure. No public exploit code or CISA KEV listing exists, but exploitation requires only knowledge of the deployed API Gateway URL and a simple HTTP request.
The graphiti FastAPI-based graph memory REST API (getzep/graphiti ≤ 0.30.2) shipped its ingest and retrieve routers with zero authentication - no API key, no token, no session check of any kind. Any unauthenticated remote caller with network access to the service can read stored knowledge-graph episodes or inject and overwrite graph data by sending ordinary HTTP requests. No public exploit code has been identified, but the barrier to exploitation is effectively zero: the attack requires only HTTP client access and knowledge of endpoint paths. No KEV listing at time of analysis.
Integer overflow in embedded-graphics ≤0.8.2 allows remote attackers to trigger memory corruption or denial-of-service by supplying a crafted `width` argument to `ImageRaw::draw_sub_image` in `src/image/image_raw.rs`. Rust's release-mode silent integer wraparound means the overflow propagates into buffer size calculations or slice indexing, potentially causing out-of-bounds memory access on the target embedded device. No vendor patch exists; the project was notified via GitHub issue #821 but has not responded. No public exploit or CISA KEV listing identified at time of analysis.
Integer overflow in the embedded-graphics Rust library (versions through 0.8.2) affects the bytes_per_row calculation inside ImageRaw::new on 32-bit compilation targets, causing incorrect buffer stride values that can lead to out-of-bounds memory access when processing crafted image data. The flaw was disclosed through GitHub issue #820, but the project had not responded with a fix at time of reporting. No public exploit code or active exploitation has been identified at time of analysis; remote reachability depends entirely on whether the consuming application ingests network-sourced image data.
Cross-tenant blob deletion and cluster-wide denial of service in Apache Storm Nimbus 3.0.0 allow authenticated topology submitters to arbitrarily delete blobstore keys owned by other tenants or permanently disable cluster leadership. Nimbus accepts `dependency_jars` and `dependency_artifacts` lists without validating their contents at submission time: during topology cleanup Nimbus deletes listed keys as itself (bypassing ACL checks), and on acquiring leadership Nimbus surrenders leadership for any topology whose dependency keys are absent - allowing a single malicious submission to loop all Nimbus daemons into perpetual leader-election, leaving the cluster inoperable. No public exploit has been identified at time of analysis, but the DoS path requires only valid topology submission credentials.
The Storm UI topology REST endpoint (`GET /api/v1/topology/{id}`) in Apache Storm 3.0.0 returned a merged Nimbus daemon and topology configuration without credential redaction, exposing ZooKeeper authentication payloads and TLS keystore/truststore passwords to any principal holding topology read-only access. This privilege gap is significant because the dedicated `getNimbusConf` API already redacted these secrets and required elevated `nimbus.users` membership, yet the topology page - which merges in strictly more daemon state - was unprotected. No public exploit is identified at time of analysis, but exploitation requires only a single authenticated HTTP GET call, making the effective complexity trivially low for any principal with read-only cluster access.
Two authorization and redaction failures in Apache Storm 3.0.0 expose credential-bearing cluster configuration to under-privileged or entirely ungated callers. The Nimbus `getNimbusConf` Thrift API serves the full daemon configuration-including ZooKeeper authentication payloads and TLS keystore and truststore passwords-after only a user-level authorization check, silently omitting the credential masking applied everywhere else in the product. The Storm UI REST endpoint `/api/v1/cluster/configuration` is worse: the missing `@AuthNimbusOp` annotation causes the authorization filter to treat the endpoint as ungated, proxying requests under the UI daemon's own principal so that any caller passing `ui.filter` receives the full unredacted configuration regardless of their privilege level. No public exploit has been identified and this CVE is not in CISA KEV.
Path traversal in Zettlab D6 Ultra NAS (versions before 1.7.0) permits a low-privileged local user to mount arbitrary host paths - including /etc/passwd and /etc/shadow - into a container with read-write access by embedding '../' sequences in Docker Compose volume declarations. An attacker with compose-file write privileges can exfiltrate password hashes or overwrite authentication files, enabling full host privilege escalation. No public exploit has been identified at time of analysis and the vulnerability is not listed in the CISA KEV catalog; version 1.7.0 resolves the issue.
Prototype pollution in hapijs/joi's messages compilation function allows network-reachable attackers to replace an object's prototype chain by supplying `__proto__` as an error-code key in custom validation messages. Affected versions are joi before 17.13.8 and before 18.2.9. Applications that pass externally-controlled data (e.g., JSON request bodies) to `.messages()` or `.prefs({ messages: ... })` can have downstream code silently broken when it relies on standard `Object.prototype` methods. No public exploit identified at time of analysis, though the patch commit ships a working proof-of-concept test file.
Authorization bypass in jaychouchannel's Tourism-Management-System allows authenticated remote users to access or modify ticket order records belonging to other users by manipulating the ID parameter in CRUD operations against MenpiaodingdanController.java. This is a classic Insecure Direct Object Reference (IDOR) pattern where the application fails to verify that the requesting user owns the referenced resource. A publicly available proof-of-concept exists (GitHub issue #9), and a patch has been committed (d44ec3aa0bd2a72c8800e3befb0a9a96a6491b86), though no versioned release is available due to the project's rolling-release approach.
Deserialization of untrusted data in the MailReceiver component of lenve vhr 1.0-SNAPSHOT enables authenticated remote attackers to manipulate application state and potentially execute arbitrary code via crafted serialized payloads. A YAKIT-based proof-of-concept exploit is publicly available on GitHub. The CVSS 4.0 score of 5.3 reflects low-privilege authentication as a prerequisite and limited per-component impact, though Java deserialization flaws can exceed their scored impact if exploitable gadget chains exist in the runtime classpath.
Authorization bypass in jaychouchannel's Tourism-Management-System (all Java/Spring Boot commits through 84d8ec384f) allows low-privileged authenticated users to invoke protected `security()` endpoints across 12 controllers by exploiting misapplied `@IgnoreAuth` annotations that instruct the AuthorizationInterceptor to skip permission enforcement entirely. The affected endpoints span chat, discussion, forum, friend, news, and store-bookmark controllers, exposing user security data and privileged actions to callers who lack the required roles. Publicly available exploit code exists via GitHub issue #12; the vendor has released a patch commit removing the offending annotations.
Improper access control in WARP-Clash-API (vvbbnn00) allows remote attackers to bypass key-based authorization in the Surge subscription handler, gaining unauthorized read and write access to subscription data. The affected function is get_surge_subscription in services/subscription.py, where manipulation of the key argument circumvents access control checks. A public proof-of-concept exploit exists on GitHub, and the product is abandoned with no upstream fix available or anticipated.
SQL injection in itsourcecode Sales and Inventory System 1.0 allows authenticated remote attackers to manipulate database queries via the `firstname` parameter in `/pages/cust_pos_trans.php`. The flaw is a classic CWE-89 unsanitized input issue in a PHP-based point-of-sale and inventory management application. A publicly available proof-of-concept exploit exists; no public exploit identified as actively exploited in the wild (not in CISA KEV).
SQL injection in itsourcecode Sales and Inventory System 1.0 allows an authenticated remote attacker to manipulate backend database queries via the `firstname` parameter in `/pages/emp_transac.php?action=add`. A publicly available proof-of-concept exploit has been released via GitHub, confirming the attack path. The vulnerability arises from unsanitized user input being passed directly to a SQL query in a PHP-based web application, enabling data extraction or database manipulation depending on database permissions.
Improper authorization in lenve vhr 1.0-SNAPSHOT allows authenticated low-privilege users to change the password of any arbitrary employee record by supplying a manipulated `hrid` parameter to the `/hr/pass` endpoint, enabling horizontal privilege escalation and account takeover across the HR application. The flaw in `HrInfoController.updatePass` accepts the target employee identifier from client input without verifying that the requesting session holds rights over that specific record - a classic IDOR pattern classified under CWE-285. A public proof-of-concept exists via GitHub advisory; the vendor was unresponsive to coordinated disclosure, and no patch has been released.
Remote code execution in cym1102 nginxWebUI up to 4.4.2 is achievable by any authenticated user via the `autoUpdate` endpoint, which accepts an arbitrary `url` parameter and downloads and executes the resulting JAR file on the host OS without domain validation. A public proof-of-concept is disclosed in GitHub issue #213, and a fix exists only as an unmerged PR (#215) - no released patched version has been confirmed. The CVSS 4.0 low-impact metrics (VC:L/VI:L/VA:L) conflict materially with the actual RCE capability.
Code injection in sanjevirau gsubs (Electron application) up to version 1.0.3 allows remote attackers to execute arbitrary code by manipulating the filename argument passed to the showQuerySuccessPage function in renderer/index.js. The vulnerability exists in the Electron renderer process and requires passive user interaction to trigger. A public proof-of-concept exploit has been published on GitHub; no patch has been released and the vendor was unresponsive during coordinated disclosure.
Authorization bypass via IDOR in jaygajera17's Spring Boot e-commerce project allows any authenticated user to overwrite an arbitrary user's profile by supplying a foreign integer `userid` to the `updateUserProfile` endpoint in `UserController.java`. The `updateUserProfile()` method never compares the supplied `userid` against the authenticated principal, so a low-privilege session can silently modify any account's username, password, or profile fields. A public Python proof-of-concept exploit is available; a corrective pull request exists but has not yet been merged.
Server-side request forgery in FlowiseAI Flowise up to version 3.0.2 allows authenticated users to redirect internal server-side HTTP requests by supplying attacker-controlled Host and X-Forwarded-Proto headers to the Evaluations endpoint, causing the server to construct the axios.post base URL from untrusted input rather than a trusted server-side value. The flaw enables the Flowise server to be leveraged as a proxy for reaching internal network resources, cloud metadata services, or adjacent microservices invisible to the external attacker. A publicly available proof-of-concept exists (GitHub issue #6687), and vendor-released patch 3.1.3 is available, though all affected versions (≤3.0.2) have been designated end-of-life by the maintainer.
Missing authorization in wxiaoqi Spring-Cloud-Platform's OnlineController allows authenticated low-privileged users to invoke getOnlineInfo without proper access checks, exposing session and online-user data beyond their intended permission scope. Affected versions are 1.0, 2.2, and 3.0 of the ace-admin module. A public proof-of-concept exploit (poc_vuln2_session_exposure.zip) has been released, though no vendor patch exists and the project has not responded to the coordinated disclosure via GitHub issue #65.
Missing authorization in wxiaoqi Spring-Cloud-Platform 3.0.1 and 3.1.0 enables remote authenticated attackers to bypass permission enforcement via the `checkUserPermission` function in the Permission Service RPC component. A proof-of-concept exploit (`poc_vuln1_failopen_authz.zip`) is publicly available on GitHub, lowering the barrier to exploitation. The vendor has not responded to the issue report filed at the project repository, leaving affected deployments without an official patch.
Unrestricted file upload in PHPGurukul Bank Locker Management System 1.0 exposes authenticated banker-role users' ability to upload arbitrary file types via the `addressproof` parameter in `/blms/banker/add-locker-form.php`, a pattern (CWE-434) that in PHP applications commonly enables remote code execution by uploading a PHP webshell to a web-accessible directory. A publicly available proof-of-concept exploit exists per the GitHub issue linked in references, elevating practical risk above what the assigned CVSS 4.0 score of 5.3 suggests. No active exploitation is confirmed in CISA KEV at time of analysis.
SQL injection in itsourcecode Sales and Inventory System 1.0 allows authenticated remote attackers to manipulate the `ID` parameter in `/pages/inv_edit1.php` to execute arbitrary SQL against the backend database. A public exploit has been disclosed via GitHub, lowering the skill barrier for opportunistic abuse. No vendor-released patch has been identified at time of analysis; the CVSS 4.0 score of 5.3 with E:P maturity reflects partial but confirmed exploitability.
Unrestricted file upload in lenve vhr 1.0-SNAPSHOT allows any authenticated user to upload arbitrary file types through the avatar upload endpoint, bypassing server-side validation entirely. The vulnerable function FastDFSUtils.upload at the /hr/userface endpoint accepts multipart file submissions without enforcing file-type or content-type restrictions, enabling upload of web shells or other malicious payloads to the FastDFS distributed storage backend. A proof-of-concept exploit is publicly available and the vendor has not responded to disclosure, leaving no official patch in place.
SQL injection in itsourcecode Sales and Inventory System 1.0 exposes the `/pages/sup_edit1.php` endpoint to database manipulation via an unsanitized `ID` parameter, enabling a low-privileged remote attacker to read or modify arbitrary backend database contents. A publicly available exploit (POC on GitHub) confirms the technique is reproducible with minimal skill. The vulnerability is no public exploit identified in CISA KEV, but the combination of network accessibility, a published POC, and trivial exploitation mechanics elevates practical risk for any instance exposed to the network.
SQL injection in GongShengyue OnlineBooks allows authenticated remote attackers to manipulate database queries through the `column` parameter of the listSplit interface in BooksServlet.java. The vulnerability enables partial data extraction, modification, and potential availability impact against the backend database. A public proof-of-concept exploit has been published on GitHub, raising near-term exploitation probability for any internet-exposed deployments.
Cross-site request forgery in Rizwan17's PHP inventory-management-system allows remote attackers to perform unauthorized state-changing actions by tricking an authenticated victim into visiting a malicious page. The vulnerability exists in includes/process.php and results in limited integrity impact - an attacker can manipulate inventory data on behalf of a logged-in user without the user's knowledge. A publicly available proof-of-concept exploit (poc_idor_updateuser.py) exists; however, no KEV listing indicates no confirmed active exploitation at time of analysis.
Memory corruption in davenardella snap7 up to version 1.4.3 allows a network-accessible, high-privileged attacker to corrupt memory by supplying a malformed DataLen argument to the TSnap7MicroClient::opUpload function in src/core/s7_micro_client.cpp. The affected library is used for communicating with Siemens S7 PLCs in industrial control environments, making the confidentiality, integrity, and availability of PLC upload sessions potentially compromised. A public proof-of-concept exists via a GitHub issue report, and the project maintainer has not yet responded or issued a patch.
Null pointer dereference in Systerel S2OPC through version 1.7.3 enables authenticated remote clients to crash the OPC-UA server by inducing heap allocation failures during subscription publish cycles. The defect in `msg_subscription_publish_bs__alloc_notification_message_items()` silently overwrites a NULL returned by a failed DataChangeNotification allocation with the result of a subsequent EventNotificationList allocation, masking the first failure and causing a deferred NULL dereference that terminates the server process. No public exploit or active exploitation has been identified; a vendor patch is available via GitLab commit 8848f051.
Authenticated SQL injection in Feng Office's reorder handler functions exposes the backend database to partial read and write access by privileged users. The vulnerability exists in `update_system_module_order` and `update_dimension_order` within `MoreController.class.php`, where the `modules` and `dims` arguments are interpolated into SQL queries without sanitization. All versions up to 3.11.13.11 are affected, a public proof-of-concept exploit exists, and the vendor has not responded to disclosure, leaving no patch available.
Cross-site scripting in the Anchor Tag Handler of Totolink A3002MU firmware version Hh-B20211125.1046 allows network-accessible attackers with low-privilege credentials to inject malicious scripts into the router's web management interface, which execute in a victim's browser upon interaction. A public proof-of-concept exploit has been published on GitHub, lowering the technical barrier for exploitation. No vendor-released patch has been identified in the available intelligence, and the firmware timestamp (November 2021) suggests this is a long-unsupported build unlikely to receive a security update.
Stored cross-site scripting in stilleshan ServerStatus 1.0 and 2.0 allows a remote authenticated attacker to inject malicious JavaScript via the `custom` argument processed in the Stats Generation component (server/src/main.cpp). When an administrator or other user subsequently views the affected stats page, the stored payload executes in their browser session, enabling session hijacking or credential theft. A public proof-of-concept exploit exists as a documented PDF on GitHub; the vendor did not respond to coordinated disclosure.
Cross-site scripting in Fengoffice Feng Office up to version 3.11.13.11 allows an authenticated low-privilege user to inject malicious JavaScript via the og_objects.name argument processed by the getTitle() function in the Task Title Output component. When a victim user subsequently views the affected task, the unsanitized payload executes in their browser. No vendor patch exists - the vendor did not respond to coordinated disclosure - and a public proof-of-concept exploit is available.
Heap buffer overflow in GPAC MP4Box's `gf_node_get_field()` function (scenegraph/base_scenegraph.c) allows local low-privilege attackers to corrupt heap memory by supplying a crafted MP4 file containing malformed MPEG-4 scenegraph data. The root cause, visible in the fix commit, is misaligned field entries across multiple MPEG-4 node struct definitions in nodes_mpeg4.h - specifically `GF_ChildNodeItem` fields misassigned between struct pairs (OctreeImage/XXParticles, SolidRep/SubdivisionSurface, PointTextureV2/SBVCAnimationV2) - causing `gf_node_get_field` to compute incorrect byte offsets and write outside intended heap allocations. A public proof-of-concept exists; no active exploitation (CISA KEV) has been confirmed.