Remote Code Execution

An unauthenticated server-side request forgery (SSRF) vulnerability exists in AVideo's Live plugin test.php endpoint that allows remote attackers to force the server to send HTTP requests to arbitrary URLs. The vulnerability affects AVideo installations with the Live plugin enabled and can be exploited to probe internal network services, access cloud metadata endpoints, and retrieve content from internal HTTP resources. A proof-of-concept has been published demonstrating localhost service enumeration, and the vulnerability requires no authentication or user interaction to exploit.

Remote code execution in PHP ffmpeg integration allows unauthenticated attackers to execute arbitrary OS commands on standalone encoder servers by bypassing incomplete input sanitization that fails to filter bash command substitution syntax. The vulnerable `sanitizeFFmpegCommand()` function strips common shell metacharacters but permits `$()` notation, which can be injected through crafted encrypted payloads and executed in a double-quoted shell context. No patch is currently available.

The Gallery plugin in AVideo contains an unauthenticated remote code execution vulnerability through CSRF-enabled PHP code injection. Attackers can exploit an eval() function that directly executes unsanitized user input by tricking an admin into visiting a malicious page, with the session cookie's SameSite=None configuration enabling cross-site request forgery. A detailed proof-of-concept exploit exists demonstrating command execution through crafted form submissions.

A critical authentication bypass and command injection vulnerability chain in AVideo's CloneSite plugin allows completely unauthenticated remote attackers to achieve full system compromise. The vulnerability affects AVideo installations with the CloneSite plugin enabled, allowing attackers to steal clone authentication keys, dump the entire database including MD5-hashed admin credentials, crack those credentials trivially, and finally execute arbitrary system commands via an rsync command injection. A detailed proof-of-concept demonstrating the complete attack chain is publicly available in the GitHub security advisory, making this an immediate exploitation risk.

ScreenToGif, a widely-used screen recording application, is vulnerable to DLL sideloading attacks through a malicious version.dll file. Versions from 2.42.1 and earlier are affected when the portable executable is run from user-writable directories, which is the primary intended use case for this application. Attackers can achieve arbitrary code execution in the user's context with high impact on confidentiality, integrity, and availability. No public patches are available at the time of disclosure, and no evidence of active exploitation (KEV status) has been reported.

libfuse versions 3.18.0 through 3.18.1 contain a use-after-free vulnerability in the io_uring subsystem that allows local attackers to crash FUSE filesystem processes or execute arbitrary code when thread creation fails under resource constraints. The flaw occurs when io_uring initialization fails (e.g., due to cgroup limits), leaving a dangling pointer in session state that is dereferenced during shutdown. Public exploit code exists for this vulnerability, and no patch is currently available.

Stack-based buffer overflow in GMT versions 6.6.0 and earlier allows local attackers to crash the application or execute arbitrary code by supplying an excessively long dataset identifier to vulnerable functions like gmt_remote_dataset_id. The vulnerability affects command-line processing of geographic data and currently lacks a public patch, leaving all affected GMT installations exposed to local exploitation.

Authenticated users can trigger a heap overflow in MariaDB 11.4 (before 11.4.10) and 11.8 (before 11.8.6) through the JSON_SCHEMA_VALID() function, causing denial of service and potentially remote code execution under specific memory layout conditions. The vulnerability requires valid database credentials and affects server availability and integrity across scope boundaries. No patch is currently available for vulnerable versions.

Precurio Intranet Portal 4.4 contains a CSRF vulnerability that allows attackers to trick authenticated users into uploading malicious files to the server, potentially leading to remote code execution with web server privileges. A public exploit is available via PacketStorm (file ID 215644), significantly lowering the barrier for exploitation. The vulnerability carries a CVSS score of 8.8 with network-based attack vector requiring only user interaction.

FastGPT versions 4.14.8.3 and below contain a critical arbitrary code execution vulnerability in the fastgpt-preview-image.yml GitHub Actions workflow that allows external contributors to execute malicious code and exfiltrate repository secrets. The vulnerability stems from unsafe use of pull_request_target with attacker-controlled Dockerfile builds that are pushed to the production container registry, enabling both direct compromise and supply chain attacks. No patch was available at the time of public disclosure, making this an unpatched remote code execution affecting the AI Agent building platform across affected versions.

WWBN AVideo open source video platform versions 25.0 and below ship with a hardcoded default administrator password ('password') in official Docker deployment files that is automatically used during installation without any forced change mechanism. Attackers can gain immediate administrative access to unpatched instances, enabling user data exposure, content manipulation, and potential remote code execution via file upload and plugin management features. The issue is compounded by weak MD5 password hashing and similarly insecure default database credentials (avideo/avideo).

SQLBot, an intelligent data query system based on large language models and RAG, contains a critical SQL injection vulnerability in the /api/v1/datasource/uploadExcel endpoint that allows authenticated users with minimal privileges to achieve remote code execution on the backend server. SQLBot versions prior to 1.7.0 are affected, and attackers can exploit unsafe concatenation of Excel sheet names into PostgreSQL table names and COPY statements to inject malicious SQL commands. The vulnerability enables arbitrary command execution as the postgres user, database takeover, and sensitive file exfiltration including /etc/passwd and /etc/shadow.

A sandbox escape vulnerability exists in Google Chrome's V8 JavaScript engine prior to version 146.0.7680.153, allowing remote attackers to execute arbitrary code within the Chrome sandbox through a crafted HTML page. This is a High severity issue affecting millions of Chrome users across Windows, macOS, and Linux platforms. The vulnerability is triggered via web-based attack vector (HTML page delivery) and does not require user interaction beyond visiting a malicious website.

Sandboxed arbitrary code execution in Google Chrome's WebAudio component (versions prior to 146.0.7680.153) can be triggered remotely through malicious HTML, requiring only user interaction. An attacker can craft a weaponized webpage to break out of the Chrome sandbox and execute arbitrary code on affected systems. This high-severity vulnerability impacts Chrome, Ubuntu, and Debian users, with patches now available.

Xerte Online Toolkits 3.14 and earlier contain an unauthenticated arbitrary file upload vulnerability allowing remote code execution with a CVSS score of 9.8. The template import functionality at /website_code/php/import/import.php lacks authentication checks, enabling attackers to upload ZIP archives containing malicious PHP files that are extracted to web-accessible directories. This is a critical severity issue with network-based attack vector requiring no privileges or user interaction, and a proof-of-concept has been published by VulnCheck.

A critical remote code execution vulnerability in SuiteCRM versions 7.15.0 and 8.9.2 allows authenticated administrators to execute arbitrary system commands through a bypass of previous security patches. This vulnerability circumvents the ModuleScanner.php security controls by exploiting improper PHP token parsing that resets security checks when encountering single-character tokens, enabling attackers to hide dangerous function calls. The vulnerability represents a direct bypass of the previously patched CVE-2024-49774 and has been assigned a CVSS score of 9.1.

An authenticated remote code execution vulnerability exists in SuiteCRM modules that allows high-privileged users to execute arbitrary code on the server. The vulnerability affects SuiteCRM versions prior to 7.15.1 and 8.9.3, and stems from improper code injection protections (CWE-94). While exploitation requires high privileges (admin-level), successful attacks grant complete control over the CRM system containing sensitive customer data.

A symlink traversal vulnerability in OpenClaw allows authenticated attackers to read and write arbitrary files on the host system through the agents.files.get and agents.files.set methods. The vulnerability affects OpenClaw versions prior to 2026.2.25 and can lead to remote code execution through strategic file overwrites. With a high CVSS score of 8.8 and an RCE tag, this represents a critical security risk for organizations using affected versions.

Remote code execution in OpenWrt's mDNS daemon (versions before 24.10.6 and 25.12.1) allows unauthenticated attackers to overflow a 46-byte stack buffer by sending malformed IPv6 PTR queries over multicast DNS on UDP port 5353. The vulnerability stems from insufficient validation of domain name length before copying to a fixed-size buffer, enabling arbitrary code execution on affected embedded devices. No patch is currently available.

A configuration injection vulnerability in Kubernetes ingress-nginx controller allows authenticated attackers to inject arbitrary nginx configuration through specially crafted Ingress annotations, leading to remote code execution with controller privileges and exposure of all cluster Secrets. The vulnerability has a high CVSS score of 8.8 and affects the ingress-nginx controller's annotation parsing mechanism. No active exploitation (not in KEV) or public POC has been reported, though the attack requires only low privileges and network access.

Remote code execution in SQLBot 1.5.0 and below allows authenticated users to inject malicious prompts through unsanitized terminology uploads, enabling attackers to manipulate the LLM into generating arbitrary PostgreSQL commands executed with database privileges. The vulnerability stems from missing permission checks on the Excel upload API combined with inadequate semantic isolation when injecting user-controlled data into the system prompt. An attacker can exploit this to achieve code execution on the database or application server running as the postgres user.

Authenticated file read vulnerability in PHP and Docker deployments allows users to exfiltrate arbitrary files from the server by exploiting insufficient path validation in the video upload endpoint, which copies attacker-specified local files to publicly accessible storage. An authenticated attacker can leverage this to read sensitive files from broad server directories including application roots, cache, and temporary locations. No patch is currently available, and the vulnerability carries a 10% exploit prediction score.

An authenticated path traversal vulnerability in Langflow's file upload functionality allows attackers to write arbitrary files anywhere on the host system, leading to remote code execution. The vulnerability affects Langflow version 1.7.3 and earlier, where the multipart upload filename bypasses security checks due to missing boundary containment in the LocalStorageService layer. A proof-of-concept exploit is publicly available demonstrating successful arbitrary file write outside the intended user directory.

Prototype pollution in the flatted npm library (versions <= 3.4.1) lets an attacker who controls input to the parse() function leak a live reference to Array.prototype into the returned object, so any later write to that property poisons the global prototype chain. The flaw stems from parse() using attacker-supplied JSON string values such as "__proto__" as raw array index keys without numeric validation. Publicly available exploit code exists (a three-line proof of concept in the GitHub Security Advisory), though EPSS is only 0.01% (2nd percentile) and it is not in CISA KEV.

Remote code execution in wgcloud version 2.3.7 and earlier allows unauthenticated attackers to execute arbitrary code through the test connection function. The vulnerability carries a critical CVSS score of 9.8 with network-based exploitation requiring no privileges or user interaction. No public exploit has been identified at time of analysis, though the EPSS score of 0.29% (52nd percentile) indicates low predicted exploitation probability despite the critical severity rating.

OpenClaw versions prior to 2026.2.22 contain an authorization bypass vulnerability in its allow-always wrapper persistence mechanism that enables remote code execution. Attackers with high privileges and user interaction can approve benign wrapped system.run commands, then subsequently execute arbitrary different payloads without requiring additional approval, compromising both gateway and node-host execution environments. A patch is available from the vendor, and this vulnerability is tagged as enabling both RCE and command injection attacks.

A remote code execution vulnerability in DedeCMS v.5.7.118 and (CVSS 9.8) that allows a remote attacker. Critical severity with potential for significant impact on affected systems.

Remote code execution in the Beijing Academy of Artificial Intelligence (BAAI) OmniGen2-RL reward server lets unauthenticated attackers run arbitrary OS commands by POSTing a malicious pickle payload to the exposed HTTP endpoint. The reward_server.py and reward_proxy.py components call pickle.loads() directly on raw, attacker-controlled request bodies, so any network-reachable instance can be fully compromised. Publicly available exploit code exists (chocapikk.com write-up) and a vendor patch (PR #139) is available, though EPSS remains low at 0.13% and there is no public exploit identified as actively exploited.

Memory exhaustion in Python's pickle deserialization allows attackers to crash applications by supplying a small malicious payload that forces allocation of gigabytes of memory through unrestricted constructor arguments in whitelisted classes. Applications using `_RestrictedUnpickler` to load untrusted pickle data are vulnerable to denial of service attacks. A patch is available.

Dynaconf, a Python configuration management library, contains a Server-Side Template Injection (SSTI) vulnerability in its @jinja resolver that allows arbitrary command execution when attackers can control configuration sources such as environment variables, .env files, or CI/CD secrets. The vulnerability affects pip package dynaconf and includes a public proof-of-concept demonstrating command execution via Jinja2 template evaluation without sandboxing. The @format resolver additionally enables object graph traversal to expose sensitive runtime data including API keys and credentials.

An unauthenticated remote code execution vulnerability exists in the mesop Python package's debugging Flask server endpoint (/exec-py) that accepts and executes arbitrary base64-encoded Python code without any authentication or validation. The vulnerability affects the mesop pip package, with a publicly disclosed proof-of-concept demonstrating trivial exploitation requiring only a single HTTP POST request. With a CVSS score of 9.8 (Critical) and detailed PoC availability, this represents an immediately exploitable vulnerability for any exposed instance.

HTSlib, a widely-used bioinformatics library for reading and writing sequence alignment formats, contains a critical buffer overflow vulnerability in its CRAM format decoder. The vulnerability exists in the `cram_byte_array_len_decode()` function which fails to validate that unpacked data matches the output buffer size, affecting HTSlib versions prior to 1.23.1, 1.22.2, and 1.21.1. An attacker can craft a malicious CRAM file that, when opened by a user, triggers either a heap or stack overflow with attacker-controlled bytes, potentially leading to arbitrary code execution, program crash, or memory corruption.

HTSlib versions prior to 1.23.1, 1.22.2, and 1.21.1 contain a heap buffer overflow vulnerability in the GZI index loading function `bgzf_index_load_hfile()`. An integer overflow during buffer allocation allows attackers to craft malicious `.gzi` files that trigger heap memory corruption, potentially leading to denial of service, data corruption, or remote code execution when a user opens the compromised file. No evidence of active exploitation in the wild has been reported, but the vulnerability is demonstrable and patch availability is confirmed.

HTSlib versions prior to 1.21.1, 1.22.2, and 1.23.1 contain an out-by-one error in the CRAM decoder's `cram_byte_array_stop_decode_char()` function that allows a single attacker-controlled byte to be written beyond the end of a heap allocation. This heap buffer overflow (CWE-122) affects bioinformatics applications using HTSlib to process CRAM-formatted DNA sequence alignment files, and could enable arbitrary code execution if exploited. No public exploit code or KEV status is currently documented, but patch availability exists for multiple stable release branches.

HTSlib contains a buffer overflow vulnerability in its CRAM format decoder affecting the VARINT and CONST encoding handlers, where incomplete context validation allows writes of up to eight bytes beyond heap allocation boundaries or into stack-allocated single-byte variables. This vulnerability affects HTSlib versions prior to 1.23.1, 1.22.2, and 1.21.1, and impacts any application using the library to process CRAM-formatted bioinformatics data files. An attacker can craft a malicious CRAM file to trigger heap or stack overflow conditions, potentially leading to denial of service, memory corruption, or arbitrary code execution when processed by a vulnerable application.

HTSlib contains a heap buffer overflow vulnerability in its CRAM decoder caused by an out-by-one error when validating feature boundaries. When a user opens a maliciously crafted CRAM file, an attacker can write one controlled byte beyond the end of a heap buffer, potentially causing application crashes, data corruption, or arbitrary code execution. Versions 1.23.1, 1.22.2, and 1.21.1 include fixes, and patches are available via the official GitHub repository.

HTSlib versions prior to 1.23.1, 1.22.2, and 1.21.1 contain a heap buffer overflow vulnerability in the cram_decode_seq() function when processing CRAM-formatted bioinformatics files with omitted sequence and quality data. An attacker can craft a malicious CRAM file that triggers an out-of-bounds read followed by an attacker-controlled single-byte write to heap memory, potentially enabling arbitrary code execution, data corruption, or denial of service when a user opens the file. No public exploit proof-of-concept has been identified, but the vulnerability is confirmed and patched by the HTSlib project.

PySpector versions 0.1.6 and earlier contain a security validation bypass in the plugin system that allows arbitrary code execution. The validate_plugin_code() function fails to detect dangerous API calls when invoked indirectly via getattr(), allowing malicious plugins to execute system commands. A public proof-of-concept exploit exists demonstrating the bypass, and while exploitation requires user interaction (installing and trusting a malicious plugin), successful exploitation grants full system access including filesystem manipulation, credential theft, and persistence mechanisms.

The NextGEN Gallery plugin for WordPress contains a Local File Inclusion vulnerability in the 'template' parameter of gallery shortcodes, affecting all versions up to and including 4.0.3. Authenticated attackers with Author-level privileges or higher can include and execute arbitrary PHP files on the server, potentially leading to remote code execution, data theft, or complete site compromise. This is a confirmed vulnerability reported by Wordfence with a high CVSS score of 8.8, though no active exploitation (KEV) status has been reported at this time.

SiYuan's Bazaar marketplace fails to sanitize package metadata (displayName, description) before rendering in the Electron desktop application, allowing stored XSS that escalates to arbitrary remote code execution. Any SiYuan user (versions ≤3.5.9) who browses the Bazaar will automatically execute attacker-controlled code with full OS-level privileges when a malicious package card renders-no installation or user interaction required. A functional proof-of-concept exists demonstrating command execution via img onerror handlers, and this vulnerability is actively tracked in GitHub's advisory database (GHSA-mvpm-v6q4-m2pf), making it a critical supply-chain risk to the SiYuan user community.

SiYuan's Bazaar (community package marketplace) fails to sanitize HTML in package README files during rendering, allowing stored XSS that escalates to remote code execution due to unsafe Electron configuration. An attacker can submit a malicious package with embedded JavaScript in the README that executes with full Node.js access when any user views the package details in the Bazaar. This affects SiYuan versions 3.5.9 and earlier across Windows, macOS, and Linux, with a CVSS score of 9.6 and multiple real-world exploitation vectors including data theft, reverse shells, and persistent backdoors.

xiaoheiFS versions up to and including 0.3.15 contain a critical remote code execution vulnerability in the plugin upload mechanism. Administrators can upload plugin ZIP files containing arbitrary binaries which the server executes without validation based on the manifest.json 'binaries' field. This allows authenticated administrators with high privileges to achieve full system compromise by uploading malicious plugin packages.

LDAP Account Manager (LAM), a web-based interface for managing LDAP directory entries, contains a local file inclusion vulnerability in its PDF export functionality that allows authenticated users to include and execute arbitrary PHP files. When chained with GHSA-88hf-2cjm-m9g8, this vulnerability enables complete remote code execution on the affected server. The vulnerability affects all versions prior to 9.5 and requires low-privilege authentication (CVSS 8.8, PR:L), tracking across 7 Ubuntu and 4 Debian releases indicates significant deployment in enterprise LDAP environments.

aaPanel v7.57.0 contains an arbitrary file upload vulnerability that allows unauthenticated or low-privileged attackers to upload malicious files and achieve remote code execution on affected systems. The vulnerability exists in the file upload functionality of the web-based server management panel, enabling attackers to bypass file type validation and execute arbitrary code with the privileges of the aaPanel process. While no CVSS score or EPSS probability is available in current sources, the Remote Code Execution impact combined with file upload attack vectors suggests critical severity; exploitation feasibility is indicated by the existence of public vulnerability research repositories.

Langflow (a visual LLM pipeline builder) contains a critical unauthenticated code execution vulnerability (CVE-2026-33017, CVSS 9.3) in the public flow build API that allows attackers to execute arbitrary Python code by supplying malicious flow data. KEV-listed with public PoC, this vulnerability enables anyone with network access to a Langflow instance to achieve server compromise through the API that builds public flows without authentication.

A critical authentication bypass vulnerability in AVideo's installation endpoint allows unauthenticated remote attackers to take over uninitialized deployments by completing the installation process with attacker-controlled credentials and database settings. The vulnerability affects AVideo installations where the configuration file does not exist (fresh deployments, container restarts without persistent storage, or re-deployments), enabling attackers to become the sole administrator with full control over the application. A detailed proof-of-concept is publicly available, and while no active exploitation has been reported in KEV, the vulnerability has a moderate EPSS score and requires only network access to exploit.

A trust boundary enforcement vulnerability in Kiro IDE allows remote attackers to execute arbitrary code when a local user opens a maliciously crafted project directory. The vulnerability affects all versions of Kiro IDE prior to 0.8.0 on all supported platforms and bypasses workspace trust protections designed to prevent unauthorized code execution. While not currently listed in CISA KEV or showing high EPSS scores, the vulnerability enables remote code execution through local user interaction.

Privilege escalation in Wazuh Manager versions 3.9.0 through 4.14.2 allows authenticated cluster nodes to achieve unauthenticated root code execution by exploiting insecure file permissions in the cluster synchronization protocol. An attacker with cluster node access can overwrite the manager's configuration file to inject malicious commands that are subsequently executed with root privileges by the logcollector service. This vulnerability affects multi-node Wazuh deployments and has no available patch.

Remote code execution in Atlassian Bamboo Data Center versions 9.6.0 through 12.1.0 allows authenticated attackers with high privileges to execute arbitrary code on affected systems with a CVSS score of 8.6. The vulnerability impacts multiple major versions with no patch currently available, requiring immediate upgrade to patched releases such as 9.6.24, 10.2.16, or 12.1.3. Organizations unable to upgrade should prioritize access controls for high-privileged accounts until remediation is possible.

A critical deserialization vulnerability in Wazuh's cluster mode allows attackers with access to any worker node to achieve remote code execution with root privileges on the master node. The vulnerability affects Wazuh versions 4.0.0 through 4.14.2 and poses severe risk to organizations using Wazuh in distributed deployments, as compromise of any single worker node can lead to full cluster takeover. While no active exploitation has been reported (not in KEV), proof-of-concept materials are publicly available via the Google Drive link in the advisory.

A critical unrestricted file upload vulnerability in Admidio's Documents & Files module allows authenticated users with upload permissions to bypass file extension restrictions by submitting an invalid CSRF token, enabling upload of PHP scripts that lead to Remote Code Execution. The vulnerability affects Admidio versions prior to the patch and has a published proof-of-concept demonstrating webshell upload and command execution. With a CVSS score of 8.8 and detailed exploitation steps available, this represents a high-priority risk for organizations using Admidio for document management.

An arbitrary file upload vulnerability in Chamilo LMS allows authenticated users with Teacher role to achieve Remote Code Execution by uploading malicious H5P packages. The flaw affects versions prior to 1.11.36 and stems from inadequate validation of H5P package contents, which only checks for h5p.json existence but fails to block .htaccess or PHP files with alternative extensions. With a CVSS score of 8.8 and high exploitation potential, attackers can upload webshells disguised as text files along with .htaccess configurations to bypass security controls.

SiYuan's mobile file tree fails to sanitize notebook names in WebSocket rename events, allowing authenticated users to inject arbitrary HTML and JavaScript that executes in other clients' browsers. When combined with Electron's insecure configuration (nodeIntegration enabled, contextIsolation disabled), this stored XSS escalates to remote code execution with full Node.js privileges on affected desktop and mobile clients. The vulnerability affects users with notebook rename permissions across Docker, Node.js, Python, and Apple platforms.

A critical remote code execution vulnerability exists in the Fields plugin for GLPI that allows authenticated users with dropdown creation privileges to execute arbitrary PHP code on the server. The vulnerability affects Fields plugin versions prior to 1.23.3 and has a CVSS score of 9.1, indicating severe impact with the ability to compromise the entire system. While no active exploitation has been reported in KEV and no public proof-of-concept is mentioned, the straightforward attack vector and high privileges requirement suggest targeted insider threat or compromised account scenarios.

A security vulnerability in renders user-supplied template content (CVSS 8.5) that allows an authenticated user with access. High severity vulnerability requiring prompt remediation. Vendor patch is available.

OIDC ID Token hash-binding bypass in the Authlib Python library (versions <= 1.6.8) lets attackers defeat at_hash and c_hash integrity checks by forging the JWT alg header with an unsupported value. The internal _verify_hash routine fails open-returning True when create_half_hash yields None for an unknown algorithm-so a forged token binds to an attacker-controlled access_token or authorization code. A detailed, working proof-of-concept against the real library exists; there is no public evidence of active exploitation, and EPSS is very low (0.02%). The flaw is fixed in Authlib 1.6.9.

CVE-2026-4276 is a security vulnerability (CVSS 7.5) that allows attackers. High severity vulnerability requiring prompt remediation.

A code injection vulnerability in SOLIDWORKS Desktop releases 2025 through 2026 allows attackers to execute arbitrary code on victim machines by tricking users into opening specially crafted files. The vulnerability requires local access and user interaction but provides complete system compromise with high impact to confidentiality, integrity, and availability (CVSS 7.8). No evidence of active exploitation or proof-of-concept code has been reported.

A DLL search order hijacking vulnerability in Thermalright TR-VISION HOME allows local attackers to execute arbitrary code with administrative privileges through DLL side-loading. The vulnerability affects versions up to 2.0.5 and occurs because the application loads DLLs using Windows' default search order without verifying integrity or signatures, allowing malicious DLLs placed in writable directories to be loaded when the application runs. No active exploitation has been reported (not in KEV), no public POC is available, and EPSS data is not yet available for this CVE.

Heap-based buffer overflow vulnerability in the DnsServer component of Tuya's arduino-TuyaOpen library (versions before 1.2.1) that allows attackers on the same LAN to execute arbitrary code on IoT/embedded devices by sending malicious DNS responses. With a CVSS score of 8.8 and tags indicating RCE capability, this represents a significant risk for connected embedded devices, though no active exploitation (not in KEV) or public PoC has been identified.

Single-byte buffer overflow vulnerability in the WiFiMulti component of arduino-TuyaOpen (versions before 1.2.1) that allows remote code execution when IoT devices connect to attacker-controlled WiFi access points. This affects Tuya's Arduino library used in smart home devices, with a CVSS score of 8.4, though the local attack vector (AV:L) suggests physical proximity is required despite the remote exploitation capability described.

Code injection in SimpleEval when objects with dangerous attrs are passed. PoC available.

A Cross-Site Scripting (XSS) vulnerability in Angular's runtime and compiler allows attackers to bypass built-in sanitization when internationalization (i18n) is enabled on security-sensitive attributes like href, src, and action. The vulnerability affects Angular versions before 19.2.20, 20.3.18, 21.2.4, and 22.0.0-next.3, enabling attackers with low privileges to execute arbitrary JavaScript in users' browsers for session hijacking, data theft, and unauthorized actions. With a CVSS score of 8.6 and no current evidence of active exploitation or public POCs, this represents a serious but not yet weaponized threat to Angular applications using i18n features with user-controlled data.

CVE-2026-3084 is an integer underflow vulnerability in GStreamer's H.266 codec parser that allows remote code execution when processing malicious media files. The vulnerability affects all versions of GStreamer (CPE indicates wildcard versioning) and can be exploited through user interaction with specially crafted H.266 video content, allowing attackers to execute arbitrary code in the context of the application. No active exploitation (not in KEV) or public POC has been reported, and the relatively high CVSS score (7.8) is tempered by the local attack vector and user interaction requirement.

CVE-2026-2921 is an integer overflow vulnerability in GStreamer's RIFF palette handling for AVI files that allows remote code execution with a CVSS score of 7.8. The vulnerability affects all versions of GStreamer (based on CPE wildcard) and requires user interaction to exploit, such as opening a malicious AVI file. No evidence of active exploitation (not in KEV), no public POC mentioned, and EPSS data not provided.

Critical out-of-bounds write vulnerability in GStreamer's rtpqdm2depay component that allows remote code execution when processing malformed X-QDM RTP payloads. The vulnerability affects all versions of GStreamer (CPE shows wildcard versioning) and requires user interaction, though attack vectors vary by implementation. With a CVSS score of 8.8 and active patch available, this represents a significant risk for applications using GStreamer for media processing.

CVE-2026-3086 is an out-of-bounds write vulnerability in GStreamer's H.266 codec parser that allows remote code execution when processing malformed APS (Adaptation Parameter Set) units. The vulnerability affects all versions of GStreamer (CPE shows wildcard versioning) and requires user interaction to exploit, such as processing a malicious H.266 video file. No evidence of active exploitation (not in KEV), no public POC, and no EPSS score available yet.

Heap-based buffer overflow vulnerability in GStreamer's rtpqdm2depay component that allows remote attackers to execute arbitrary code when processing malformed X-QDM RTP payloads. The vulnerability affects all versions of GStreamer (CPE indicates no version restrictions) and requires user interaction to exploit, though attack vectors may vary based on implementation. No active exploitation is known (not in KEV), and no EPSS score is available to assess real-world exploitation probability.

Heap-based buffer overflow vulnerability in the GStreamer multimedia framework's JPEG parser that allows remote code execution when processing malicious Huffman tables. The vulnerability affects all versions of GStreamer (CPE shows wildcard versioning) and requires user interaction to exploit, with a CVSS score of 7.8. No active exploitation in the wild has been reported (not in KEV), and no EPSS data is available.

Stack-based buffer overflow in GStreamer's H.266 codec parser that allows remote code execution when processing malicious video files. The vulnerability affects all versions of GStreamer (CPE shows wildcard versioning) and requires user interaction to trigger, such as opening a malicious media file. No active exploitation (not in KEV) or public PoC has been reported, with EPSS data unavailable.

CVE-2026-2923 is an out-of-bounds write vulnerability in GStreamer's DVB Subtitles handling that allows remote code execution when processing malformed subtitle coordinates. This vulnerability affects all versions of GStreamer (CPE indicates no version restrictions) and requires user interaction to exploit, though attack vectors may vary by implementation. No evidence of active exploitation (not in KEV), no public POC available, and no EPSS data provided.

Critical remote code execution vulnerability in GStreamer's RealMedia demuxer component, allowing attackers to execute arbitrary code via malformed video packets that trigger an out-of-bounds write. The vulnerability affects all versions of GStreamer (CPE indicates wildcard versioning) and requires user interaction to process malicious media files. While no active exploitation is reported (not in KEV), the availability of a vendor patch and ZDI advisory suggests this vulnerability has been responsibly disclosed and addressed.

Heap-based buffer overflow vulnerability in GStreamer's ASF Demuxer component that allows remote attackers to execute arbitrary code when processing malicious ASF media files. The vulnerability requires user interaction (opening/processing a malicious file) and affects all versions of GStreamer based on the CPE data. No evidence of active exploitation (not in KEV) or public proof-of-concept exists, though Zero Day Initiative tracked it as ZDI-CAN-28843.

Critical path traversal vulnerability in Unraid's update.php file that allows authenticated remote attackers to execute arbitrary code as root. The vulnerability affects all versions of Unraid (per CPE data) and was discovered by Zero Day Initiative (ZDI-CAN-28951). With a CVSS score of 8.8 and requiring only low privileges, this represents a severe risk for Unraid installations.

CVE-2026-3562 is an authentication bypass vulnerability in Philips Hue Bridge's HAP (HomeKit Accessory Protocol) implementation, specifically within the ed25519_sign_open function that fails to properly verify Ed25519 cryptographic signatures. Network-adjacent attackers can exploit this flaw without authentication to execute arbitrary code on affected Hue Bridge installations. The CVSS score of 6.3 reflects moderate severity with local network access requirements, though the authentication bypass nature elevates real-world risk for smart home environments.

Heap-based buffer overflow vulnerability in Philips Hue Bridge devices that allows network-adjacent attackers to execute arbitrary code through malformed PUT requests to the HomeKit Accessory Protocol (HAP) characteristics endpoint. While authentication is normally required, the advisory notes the authentication mechanism can be bypassed, effectively allowing unauthenticated remote code execution. No EPSS score or KEV listing is available, suggesting this is not currently being exploited in the wild.

Heap-based buffer overflow vulnerability in Philips Hue Bridge's HomeKit implementation that allows unauthenticated network-adjacent attackers to execute arbitrary code. The vulnerability affects all versions of Philips Hue Bridge (CPE indicates no version restrictions) through the hk_hap_pair_storage_put function on TCP port 8080. No EPSS data or KEV listing is available, and while ZDI has published an advisory, no public POC or active exploitation has been reported.

Heap-based buffer overflow vulnerability in Philips Hue Bridge devices that allows network-adjacent attackers with authentication (which can be bypassed) to achieve remote code execution as root. The vulnerability affects the HomeKit Accessory Protocol (HAP) implementation on TCP port 8080 and has a high CVSS score of 8.0, though no active exploitation or public PoC has been reported.

Critical heap-based buffer overflow vulnerability in Philips Hue Bridge's HomeKit implementation that allows network-adjacent attackers to execute arbitrary code without authentication. The vulnerability affects all versions of Philips Hue Bridge (CPE indicates no version restriction) and stems from improper input validation in the hk_hap_pair_storage_put function. No active exploitation (not in KEV) or EPSS score is reported, but the high CVSS score (8.8) and RCE capability make this a significant threat for local network attackers.

Heap-based buffer overflow vulnerability in the Philips Hue Bridge's Zigbee stack that allows network-adjacent attackers to execute arbitrary code when users initiate device pairing. The vulnerability affects all versions of Philips Hue Bridge and has a CVSS score of 8.0, requiring physical proximity and user interaction to exploit. No EPSS data or KEV listing is available, suggesting this is not actively exploited in the wild.

XSS in AnythingLLM 1.11.1 and earlier.

Chrome's V8 JavaScript engine contains an inappropriate implementation (CVE-2026-3910, CVSS 8.8) that allows remote attackers to execute arbitrary code within the browser sandbox via crafted HTML pages. KEV-listed with public PoC, this V8 vulnerability affects all Chromium-based browsers and enables drive-by exploitation through any web page containing malicious JavaScript.

SQL injection in OneUptime telemetry API before 10.0.23.

Stack overflow in HMS Networks Ewon Flexy/Cosy+ firmware.

HMS Networks' industrial IoT gateways (Ewon Flexy and Cosy+) contain a command injection vulnerability that allows authenticated attackers to execute arbitrary OS commands remotely. This affects Flexy devices before firmware 15.0s4 and Cosy+ devices before 22.1s6 (22.x branch) or 23.0s3 (23.x branch). With a CVSS score of 8.8 but low EPSS of 0.06%, this vulnerability requires valid credentials but enables full system compromise.

Unauthenticated attackers can trigger out-of-bounds memory access in the web interface of multiple Omada switches through improper input validation, potentially achieving remote code execution or causing denial-of-service. Affected products include Sg2005p PD 1.x, Sg2008 4.2x/4.3x, and Sg2008p 3.2x/3.3x, which require only network access to the vulnerable interface. A patch is available to address this high-severity vulnerability (CVSS 7.7).

create_function() sandbox bypass via unsanitized args passed to Function constructor. PoC available.

SandboxJS sandbox escape before 0.8.34 via Function access through arrays. CVSS 10.0.

Remote code execution in Clasp versions below 3.2.0 allows unauthenticated attackers to execute arbitrary code by uploading Google Apps Script projects with specially crafted filenames that exploit path traversal weaknesses. Public exploit code exists for this vulnerability, and no patch is currently available. The attack requires minimal user interaction and affects Google's Clasp tooling across all configurations.

Memory corruption vulnerability in all versions of Digilent DASYLab data acquisition software that occurs when processing maliciously crafted files, potentially allowing attackers to leak sensitive information or execute arbitrary code. The vulnerability requires user interaction (opening a malicious file) and has a CVSS score of 7.8, with no current evidence of active exploitation or public proof-of-concept code.