Remote Code Execution

Remote code execution via stored XSS in Notesnook Web Clipper affects all platforms prior to version 3.3.11 (Web/Desktop) and 3.3.17 (Android/iOS). Attackers can inject malicious HTML attributes into clipped web content that execute JavaScript in the application's security context when victims open the clip. On Electron desktop builds, unsafe Node.js integration (nodeIntegration: true, contextIsolation: false) escalates this XSS to full RCE with system-level access. CVSS 9.6 (Critical) reflects network-based attack requiring no authentication but user interaction. No public exploit identified at time of analysis, though attack methodology is detailed in vendor advisory.

Arbitrary code execution in Adobe Substance3D Stager 3.1.7 and earlier allows local attackers to execute malicious code with user privileges through specially crafted files. Exploitation requires social engineering to trick users into opening weaponized Stager project files. No public exploit identified at time of analysis, though the use-after-free vulnerability class is well-understood and exploitable. CVSS 7.8 (High) reflects significant impact if exploited, though local attack vector and user interaction requirement reduce immediate risk compared to remotely exploitable flaws.

JavaScript code injection in Windmill's NativeTS executor allows workspace administrators to achieve remote code execution by embedding malicious payloads in environment variable values. The vulnerability (CWE-94) stems from improper sanitization of single quotes when interpolating workspace environment variables into JavaScript string literals, enabling arbitrary code execution in all NativeTS scripts within the affected workspace. Windmill versions prior to 1.664.0 are affected. CVSS 7.3 reflects high confidentiality, integrity, and availability impact, though exploitation requires high privileges (workspace admin role). Publicly available exploit code exists, though no confirmed active exploitation (CISA KEV) at time of analysis.

Remote code execution in gematik Authenticator (macOS) versions 4.12.0 through 4.15.x enables malicious file-triggered command injection when victims open crafted documents. This CWE-78 OS command injection flaw requires no authentication but depends on user interaction (CVSS:3.1/AV:L/AC:L/PR:N/UI:R). No public exploit identified at time of analysis, though EPSS data not available. The authenticator serves German digital health applications, making this a high-impact target for healthcare sector attacks.

Remote code execution in nanobot personal AI assistant (versions prior to 0.1.6) allows unauthenticated attackers to execute arbitrary LLM instructions and system tools via malicious email content. The vulnerability exploits the email channel processing module's lack of input validation, enabling zero-click, indirect prompt injection attacks without bot owner interaction. Publicly available exploit code exists. With CVSS 8.9 (Critical) and network-accessible attack vector requiring no privileges, this represents a severe security risk for deployed nanobot instances monitoring email.

Ruby Language Server (ruby-lsp) allows arbitrary code execution when opening malicious projects. The vulnerability exploits unsanitized interpolation of the rubyLsp.branch workspace setting into a generated Gemfile, enabling attackers to embed malicious Ruby code in .vscode/settings.json that executes when users open and trust the workspace. Affects ruby-lsp gem < 0.26.9 and VS Code extension < 0.10.2. No active exploitation or public POC currently identified at time of analysis, but the attack requires only social engineering to trick developers into opening a crafted repository.

Unauthenticated remote cache poisoning in nektos/act (GitHub Actions local runner) enables arbitrary code execution by exposing the built-in actions/cache server on all network interfaces without authentication. Attackers who can reach the cache server-including from the public internet if exposed-can inject malicious cache entries with predictable keys, leading to remote code execution within Docker containers running GitHub Actions workflows. No public exploit identified at time of analysis, though EPSS data unavailable. Vendor-released patch available in act v0.2.86.

Command injection in nektos/act (GitHub Actions local runner) allows attackers to execute arbitrary code by embedding deprecated workflow commands in untrusted input. Act versions prior to 0.2.86 unconditionally process ::set-env:: and ::add-path:: commands that GitHub Actions disabled in 2020, enabling PATH hijacking and environment variable injection when workflows echo PR titles, branch names, or commit messages. Publicly available exploit code exists with working proof-of-concept demonstrating NODE_OPTIONS and LD_PRELOAD injection vectors. This creates a critical supply chain risk where workflows safe on GitHub Actions become exploitable when developers test them locally with act.

KQL injection in adx-mcp-server Python package allows authenticated attackers to execute arbitrary Kusto queries against Azure Data Explorer clusters. Three MCP tool handlers (get_table_schema, sample_table_data, get_table_details) unsafely interpolate the table_name parameter into query strings via f-strings, enabling data exfiltration from arbitrary tables, execution of management commands, and potential table drops. Vendor-released patch available (commit 0abe0ee). No public exploit identified at time of analysis, though proof-of-concept code exists in the security advisory demonstrating injection via comment-based bypass and newline-separated commands. Affects adx-mcp-server ≤ commit 48b2933.

Remote code execution in Handlebars templating engine (npm package) allows unauthenticated network attackers to execute arbitrary server-side commands by exploiting dynamic partial resolution logic. Affected versions include all releases prior to v4.7.9. Attack requires the adversary to control context data passed to templates that use dynamic partial lookups. A proof-of-concept exploit demonstrates arbitrary code execution and is publicly documented. CVSS score of 8.1 reflects high complexity due to the need for specific template patterns and attacker-controlled context values.

Remote code execution in Handlebars templating engine (npm package) allows unauthenticated attackers to execute arbitrary JavaScript on Node.js servers by exploiting the @partial-block mechanism when combined with vulnerable helper functions. The attack overwrites @partial-block with a malicious Handlebars AST that is dynamically compiled and executed during template rendering. A working proof-of-concept exists demonstrating exploitation via the commonly-used handlebars-helpers package. Vendor-released patch is available in Handlebars version 4.7.9.

Remote code execution in Handlebars.js npm package allows unauthenticated attackers to execute arbitrary JavaScript on Node.js servers by injecting malicious payloads through crafted AST objects passed to Handlebars.compile(). The vulnerability (CWE-94 code injection) affects applications that accept user-controlled JSON and deserialize it as template input. A detailed proof-of-concept exploit demonstrates command execution via process.getBuiltinModule. Vendor patch is available in version 4.7.9 per GitHub advisory GHSA-2w6w-674q-4c4q. CVSS score 9.8 (Critical) reflects network-accessible attack requiring no privileges or user interaction.

Remote code execution is achievable in Grafana installations through a chained attack combining SQL Expressions with a Grafana Enterprise plugin, affecting both open-source and Enterprise deployments. The vulnerability requires high-privilege authenticated access (PR:H) but enables cross-scope impact with complete system compromise once exploited. Only instances with the sqlExpressions feature toggle enabled are vulnerable, though Grafana recommends all users update to prevent future exploitation paths using this attack vector. No public exploit identified at time of analysis, and authentication as a high-privilege user is required per CVSS vector.

Remote code execution in Bludit CMS versions prior to 3.18.4 allows authenticated attackers holding valid API tokens to upload and execute arbitrary files through the API plugin's unrestricted file upload mechanism. The vulnerability has a CVSS 4.0 score of 8.7 with network attack vector and low complexity, requires authenticated access (PR:L), and was reported by CERT-PL. No public exploit identified at time of analysis, though the technical details are publicly disclosed.

BUFFALO Wi-Fi router products allow remote code execution through a code injection vulnerability requiring user interaction. An unauthenticated attacker (CVSS PR:N) can execute arbitrary code on affected devices with high impact to confidentiality, integrity, and availability (CVSS 8.8). The vulnerability was disclosed through JVN and BUFFALO's official advisory, with no public exploit identified at time of analysis.

Spring AI versions 1.0.0 through 1.0.4 and 1.1.0 through 1.1.3 allow unauthenticated remote code execution through Spring Expression Language (SpEL) injection in the SimpleVectorStore component when user-supplied input is incorporated into filter expression keys. This critical vulnerability (CVSS 9.8) enables attackers to execute arbitrary code without authentication on applications using SimpleVectorStore with untrusted filter input. No public exploit identified at time of analysis, though the attack complexity is low and requires no user interaction according to the CVSS vector (AV:N/AC:L/PR:N/UI:N).

Remote Code Execution and Arbitrary File Read in Metabase Enterprise Edition allows authenticated administrators to execute arbitrary code and read sensitive files via malicious serialization archives. Affected versions span at least 1.47 through 1.59.3, with patches released in versions 1.54.22, 1.55.22, 1.56.22, 1.57.16, 1.58.10, and 1.59.4. The vulnerability exploits the POST /api/ee/serialization/import endpoint by injecting INIT properties into H2 JDBC specifications within crafted serialization archives, triggering arbitrary SQL execution during database synchronization. Authentication as an admin is required (CVSS PR:H), and the vulnerability has been confirmed exploitable on Metabase Cloud infrastructure.

Remote code execution is possible in vLLM inference and serving engine versions 0.10.1 through 0.17.x due to hardcoded trust_remote_code=True settings in two model implementation files that override users' explicit --trust-remote-code=False security configuration. Attackers can exploit this by hosting malicious model repositories that execute arbitrary code when loaded by vLLM, even when users have intentionally disabled remote code trust for security. Version 0.18.0 patches this vulnerability, with no public exploit identified at time of analysis and a CVSS score of 8.8 requiring user interaction to trigger.

CodeRider-Kilo's command auto-approval module fails to correctly parse Windows CMD escape sequences (^), allowing attackers to bypass its Git command whitelist and achieve arbitrary remote code execution. The vulnerability exploits a mismatch between the Unix-based shell-quote parser used for validation and the actual Windows CMD interpreter behavior, enabling attackers to inject malicious commands through crafted payloads such as git log ^" & malicious_command ^". No public exploit code or active exploitation has been confirmed at the time of analysis.

Prompt injection attacks in AI Code's automatic command execution feature allow remote attackers to bypass the model-based safety classification system and achieve arbitrary command execution without user approval. The vulnerability affects AI Code extensions (notably the Claude Dev China variant available on the Visual Studio Code Marketplace) by exploiting the model's susceptibility to crafted prompts that misclassify destructive commands as safe. No public exploit code or confirmed active exploitation has been identified at the time of analysis, but the attack requires no authentication and can be triggered by any user with access to the extension's command execution interface.

A command injection vulnerability in command auto-approval module in Axon Code (CVSS 9.8). Critical severity with potential for significant impact on affected systems.

Remote code execution in happy-dom ECMAScriptModuleCompiler allows attackers to execute arbitrary OS commands by injecting JavaScript expressions via backtick-bypassed export declarations. Affects npm package happy-dom versions 15.10.0 through 20.8.7. Exploitation requires JavaScript evaluation to be explicitly enabled (disabled by default since v20.0.0). A public proof-of-concept exists, but no active exploitation or CISA KEV listing; EPSS probability is low (0.07%).

Unauthenticated remote code execution as root is possible in thingino-firmware through the WiFi captive portal CGI script due to command injection in query and POST parameter parsing. Attackers on the adjacent network (AV:A) can inject arbitrary commands through unsanitized HTTP parameter names, enabling full device takeover including root password reset and SSH key manipulation for persistent access. No public exploit is identified at time of analysis, though VulnCheck has published an advisory detailing the vulnerability mechanics.

Prototype pollution in convict npm package version 6.2.4 allows attackers to bypass previous security fixes and pollute Object.prototype through crafted input that manipulates String.prototype.startsWith. The vulnerability affects applications processing untrusted input via convict.set() and can lead to authentication bypass, denial of service, or remote code execution if polluted properties reach dangerous sinks like eval or child_process. A working proof-of-concept exploit demonstrating the bypass technique exists in the advisory.

Prototype pollution in Mozilla's node-convict configuration library allows attackers to inject properties into Object.prototype via two unguarded code paths: config.load()/loadFile() methods that fail to filter forbidden keys during recursive merge operations, and schema initialization accepting constructor.prototype.* keys during default-value propagation. Applications using node-convict (pkg:npm/convict) that process untrusted configuration data face impacts ranging from authentication bypass to remote code execution depending on how polluted properties propagate through the application. This represents an incomplete fix for prior prototype pollution issues (GHSA-44fc-8fm5-q62h), with no public exploit identified at time of analysis.

CVE-2026-33870 is a security vulnerability (CVSS 7.5). High severity vulnerability requiring prompt remediation.

The Zen C compiler (versions prior to 0.4.4) crashes or enables arbitrary code execution when processing maliciously crafted .zc source files containing excessively long identifiers for structs, functions, or traits, triggering a stack-based buffer overflow (CWE-121). A proof-of-concept exploit exists per SSVC assessment, though attack complexity remains moderate as it requires local access and user interaction (CVSS:3.1/AV:L/AC:L/PR:N/UI:R). Vendor-released patch: version 0.4.4.

Langflow's Agentic Assistant feature executes LLM-generated Python code server-side during component validation, enabling arbitrary code execution when attackers can influence model outputs. The vulnerability affects the pip package 'langflow' and exists in endpoints /assist and streaming paths that invoke exec() on dynamically generated component code. A proof-of-concept exists demonstrating the execution chain from user input through validation to code execution. Authentication requirements depend on deployment configuration, with AUTO_LOGIN=true defaults potentially widening exposure. No public exploit identified at time of analysis beyond the documented PoC, though the technical details and code references provide a complete exploitation blueprint.

Remote code execution is possible in DataDog's dd-trace-java agent versions prior to 1.60.3 when running on JDK 16 or earlier with exposed JMX/RMI ports. The vulnerability stems from unsafe deserialization in the RMI instrumentation's custom endpoint, allowing network-accessible attackers to execute arbitrary code if gadget-chain libraries exist on the classpath. Vendor-released patch: version 1.60.3. No public exploit identified at time of analysis, though the issue was responsibly disclosed through DataDog's bug bounty program by Mohamed Amine ait Ouchebou.

Stack-based buffer overflow in EVerest EV charging software allows unauthenticated local attackers to execute arbitrary code via overly long CAN interface names during initialization. The vulnerability (CWE-121) affects everest-core versions prior to 2026.02.0 with CVSS 8.4 (High severity). Proof-of-concept exploit code exists according to SSVC assessment, and the flaw triggers before privilege checks, enabling attack with no user privileges required. The vulnerability is tracked as EUVD-2026-16199 by ENISA.

Remote code execution vulnerability in EVerest electric vehicle charging software stack allows adjacent network attackers to execute arbitrary code by sending malformed SLAC protocol frames. EVerest-core versions prior to 2026.02.0 are affected due to a stack buffer overflow in HomeplugMessage::setup_payload that trusts an attacker-controlled length parameter in release builds. SSVC analysis indicates proof-of-concept exploit code exists, though the vulnerability is not automatable and requires adjacent network access (CVSS 8.8, AV:A).

Stack-based buffer overflow in EVerest EV charging software stack enables local code execution when processing certificate filenames of exactly 100 characters due to off-by-one boundary check error in IsoMux component. EVerest-core versions prior to 2026.02.0 are affected (CPE cpe:2.3:a:everest:everest-core). The vulnerability has a CVSS score of 8.4 with local attack vector and no privilege requirements (AV:L/PR:N), allowing unauthenticated local attackers to achieve code execution. No public exploit identified at time of analysis, though technical details are available in GitHub security advisory GHSA-cpqf-mcqc-783m.

Remote code execution is achievable in Red Hat Foreman and Satellite 6 via command injection in the WebSocket proxy implementation when users access VM VNC console functionality. An attacker controlling a malicious compute resource server can inject unsanitized hostname values into shell commands, compromising the Foreman server and potentially the entire managed infrastructure. A proof-of-concept exploit exists according to SSVC data, elevating real-world risk despite requiring low-privileged authentication and user interaction.

Remote code execution in plank/laravel-mediable PHP package through version 6.4.0 allows unauthenticated attackers to upload executable PHP files disguised with benign MIME types, achieving arbitrary code execution when files land in web-accessible directories. EPSS score of 0.39% (60th percentile) indicates low observed exploitation probability, though SSVC analysis confirms the vulnerability is automatable with total technical impact. No vendor-released patch identified at time of analysis despite coordinated disclosure attempts.

BentoML, a Python framework for ML model serving, contains a command injection vulnerability in the docker.system_packages configuration field of bentofile.yaml files. The vulnerability affects all versions supporting this feature (confirmed in version 1.4.36) and allows attackers to execute arbitrary commands during the Docker image build process (bentoml containerize). This is a high-severity supply chain risk with a CVSS score of 7.8, requiring user interaction to trigger but achieving full command execution as root during container builds.

RATOC RAID Monitoring Manager for Windows contains an insecure directory permissions vulnerability when the installation folder is customized to a non-default location. The installer fails to properly set access control lists (ACLs) on custom installation directories, allowing non-administrative users to modify folder contents and execute arbitrary code with SYSTEM privileges. With a CVSS 4.0 score of 8.5, this represents a high-severity local privilege escalation vulnerability affecting Windows systems where this RAID management software is installed.

RATOC RAID Monitoring Manager for Windows contains a DLL hijacking vulnerability in its installer that loads DLLs from the current directory without proper path validation. If an attacker can place a malicious DLL in the directory where a user runs the installer, arbitrary code can be executed with administrator privileges. The vulnerability has a CVSS score of 8.4 with local attack vector requiring user interaction, and has been publicly disclosed through JPCERT coordination with vendor advisory available.

Remote code execution in FreeBSD kernel's RPCSEC_GSS implementation (kgssapi.ko) and userspace RPC servers (librpcgss_sec) allows low-privileged authenticated attackers to execute arbitrary code via crafted network packets. Affects FreeBSD 13.5, 14.3, 14.4, and 15.0 branches. Publicly available exploit code exists (GitHub/Gist), though EPSS probability remains low (0.05%, 16th percentile) and CISA has not listed this in KEV, suggesting limited observed exploitation despite high CVSS 8.8 score and total technical impact per SSVC framework.

The GREEN HOUSE CO., LTD. Digital Photo Frame GH-WDF10A contains active debug code that allows unauthenticated local attackers to read or write arbitrary files and execute commands with root privileges. This vulnerability affects all versions of the GH-WDF10A model and represents a critical local privilege escalation risk for any user with physical or network access to the device. While the CVSS score of 6.8 reflects medium severity due to the physical access requirement, the ability to achieve root code execution makes this a significant concern for device owners and enterprise deployments.

The Frontend Admin by DynamiApps plugin for WordPress contains a PHP Object Injection vulnerability affecting all versions up to and including 3.28.31. Authenticated attackers with Editor-level privileges or higher can exploit unsafe deserialization of the 'post_content' field in admin_form posts to inject malicious PHP objects and achieve remote code execution through available POP chains. This represents a critical risk for WordPress sites using this plugin with elevated user accounts.

Saloon PHP library versions prior to 4.0.0 contain a PHP object injection vulnerability in the AccessTokenAuthenticator::unserialize() method, which unsafely deserializes OAuth token state using unserialize() with allowed_classes set to true. An attacker who can control the serialized token string-such as by overwriting a cached token file or injecting malicious data-can supply a crafted serialized gadget object that executes arbitrary code through PHP magic methods during deserialization. In environments with common dependencies like Monolog present, this vulnerability can be reliably chained to achieve remote code execution (RCE), making it a critical threat to any API integration or SDK built on vulnerable Saloon versions.

Remote code execution in Daylight Studio FuelCMS v1.5.2 through the /parser/dwoo component enables unauthenticated attackers to execute arbitrary PHP code via specially crafted input. The vulnerability exploits insufficient input validation in the Dwoo template engine integration, allowing direct PHP code injection. Attack complexity appears low given the public references to exploitation techniques in the provided pentest-tools PDF, though no formal CVSS scoring or CISA KEV confirmation is available to assess real-world exploitation prevalence.

Authenticated attackers with Subscriber-level access can delete arbitrary files on WordPress servers running WP Job Portal plugin versions up to 2.4.9, enabling remote code execution by removing critical files like wp-config.php. The vulnerability stems from insufficient file path validation in the removeFileCustom function. EPSS exploitation probability is 0.25% (48th percentile), indicating low predicted real-world exploitation likelihood, though the CVSS score of 8.8 reflects high potential impact when successfully exploited. No public exploit identified at time of analysis.

A deserialization vulnerability in OpenTelemetry Java instrumentation versions prior to 2.26.1 allows remote code execution when the RMI instrumentation endpoint processes untrusted data without serialization filters. The vulnerability affects applications using the OpenTelemetry Java agent with network-reachable RMI/JMX endpoints and gadget-chain-compatible libraries on the classpath. This was responsibly disclosed in coordination with Datadog, and a patch is available in version 2.26.1.

The code16/sharp Laravel admin panel package contains a critical file upload vulnerability that allows authenticated users to bypass all file type restrictions by manipulating client-controlled validation rules. Affected versions prior to 9.20.0 accept a user-supplied validation_rule parameter that is passed directly to Laravel's validator, enabling attackers to upload arbitrary files including PHP webshells. With a CVSS score of 8.8, this vulnerability can lead to Remote Code Execution when the storage disk is publicly accessible, though default configurations provide some protection against direct execution.

A prototype pollution vulnerability in the XML and GSuiteAdmin nodes of n8n workflow automation platform allows authenticated users with workflow creation or modification permissions to achieve remote code execution. Versions prior to 2.14.1, 2.13.3, and 1.123.27 are affected. The CVSS score of 9.4 (Critical) reflects network-based exploitation with low complexity requiring only low-level authentication, though no current KEV listing or public POC availability is indicated in the provided intelligence.

An authenticated user with workflow creation or modification privileges in n8n workflow automation platform can exploit the Merge node's 'Combine by SQL' mode to read arbitrary local files on the n8n host and achieve remote code execution. n8n versions prior to 2.14.1, 2.13.3, and 1.123.26 are affected. The vulnerability carries a CVSS 4.0 score of 9.4 (Critical) due to insufficient sandbox restrictions in the AlaSQL component, allowing SQL injection-style attacks against the host system. No public proof-of-concept or active exploitation (KEV) status has been reported at this time.

A Code Injection vulnerability (CWE-94) exists in Nelio AB Testing WordPress plugin through version 8.2.7 that allows attackers to execute arbitrary code on affected installations. The vulnerability affects the Nelio Software product across all versions up to and including 8.2.7, potentially enabling remote code execution (RCE). This is a critical severity issue as it permits unauthenticated or low-privilege attackers to gain complete control over WordPress sites running the vulnerable plugin.

A Code Injection vulnerability (CWE-94) exists in JetFormBuilder versions up to and including 3.5.6.1, allowing attackers to inject and execute arbitrary code within the application context. The vulnerability affects the JetFormBuilder plugin for WordPress across all versions through 3.5.6.1, and an attacker can leverage this to achieve Remote Code Execution (RCE) by injecting malicious code through form-processing mechanisms. Patchstack has documented this vulnerability with an assigned EUVD ID (EUVD-2026-15889), and while a CVSS score has not been formally assigned, the RCE classification indicates critical severity.

Total Poll Lite, a WordPress plugin, contains an improper code injection vulnerability (CWE-94) that allows remote code inclusion and execution. All versions up to and including 4.12.0 are affected. An attacker can exploit this vulnerability to achieve remote code execution (RCE) on WordPress installations running the vulnerable plugin, potentially gaining full control of the affected web application.

A Code Injection vulnerability (CWE-94) exists in the Jonathan Daggerhart Widget Wrangler WordPress plugin through version 2.3.9, allowing unauthenticated attackers to execute arbitrary code on affected installations. This Remote Code Execution (RCE) vulnerability enables complete server compromise and data exfiltration. Active exploitation has been documented by Patchstack, indicating this is a practical, real-world threat requiring immediate patching.

A Code Injection vulnerability exists in the Themeisle Woody ad snippets plugin (insert-php) through version 2.7.1 that allows unauthenticated attackers to execute arbitrary PHP code on affected WordPress installations. The vulnerability stems from improper control of code generation, classified as CWE-94, enabling remote code execution (RCE). Patchstack has documented this issue, and affected installations should be patched immediately as the attack vector appears to be network-accessible with low complexity.

The Post Snippets WordPress plugin versions up to and including 4.0.12 contain an improper code generation vulnerability (CWE-94) that enables remote code injection and execution. An attacker can exploit this flaw to execute arbitrary code on affected WordPress installations, potentially leading to complete site compromise. The vulnerability has been publicly documented by Patchstack with available references, and the attack vector appears to be network-based without requiring high privileges.

A DLL hijacking vulnerability exists in the installer for OM Workspace (Windows Edition) Ver 2.4 and earlier, allowing local attackers to execute arbitrary code with the privileges of the user running the installer. The vulnerability is reported by JPCERT and affects software from OM Digital Solutions Corporation. With a CVSS score of 7.8 (High), the vulnerability requires local access and user interaction but no special privileges, making it a moderate real-world risk for targeted attacks during software installation.

A remote code execution vulnerability (CVSS 8.4). High severity vulnerability requiring prompt remediation.

Thumbler through version 1.1.2 contains an OS command injection vulnerability in the thumbnail() function where user-supplied input from the input, output, time, or size parameters is directly concatenated into shell commands executed via Node.js child_process.exec() without sanitization or escaping. This allows unauthenticated attackers to execute arbitrary operating system commands with the privileges of the application process. A proof-of-concept has been documented in public repositories, making this vulnerability immediately actionable for exploitation.

The textract library through version 2.5.0 contains an OS command injection vulnerability in its file extraction modules that allows attackers to execute arbitrary operating system commands by crafting malicious filenames. The vulnerability affects multiple extractors (doc.js, rtf.js, dxf.js, images.js, and util.js) where user-supplied file paths are passed directly to child_process.exec() without adequate sanitization. An attacker can exploit this by uploading or referencing files with specially crafted names containing shell metacharacters, leading to complete system compromise with the privileges of the process running textract.

The pdf-image npm package through version 2.0.0 contains an OS command injection vulnerability in the pdfFilePath parameter. Attackers can exploit this remotely without authentication by injecting malicious commands through file path inputs that are passed unsafely to shell commands via child_process.exec(). A proof-of-concept exploit is publicly available on GitHub (zebbernCVE/CVE-2026-26830), significantly increasing exploitation risk.

NVIDIA NeMo Framework contains an insecure deserialization vulnerability (CWE-502) that allows authenticated local attackers to execute arbitrary code. The vulnerability affects NVIDIA NeMo Framework installations and can lead to code execution, privilege escalation, information disclosure, and data tampering. According to CISA's SSVC framework, there is currently no evidence of active exploitation in the wild, and the attack is not automatable, though technical impact is rated as total.

NVIDIA NeMo Framework contains a remote code execution vulnerability in its checkpoint loading mechanism caused by insecure deserialization (CWE-502). Attackers with local access and low privileges can exploit this to achieve code execution, privilege escalation, information disclosure, and data tampering with high impact on confidentiality, integrity, and availability. According to SSVC framework, there is currently no observed exploitation in the wild, though the technical impact is rated as total.

NVIDIA Model Optimizer for Windows and Linux contains an unsafe deserialization vulnerability in its ONNX quantization feature that allows attackers to execute arbitrary code by providing a malicious input file. Users who process untrusted ONNX model files are at risk of complete system compromise, including code execution, privilege escalation, data tampering, and information disclosure. There is no current evidence of active exploitation (not in CISA KEV) or public proof-of-concept availability.

NVIDIA Megatron-LM contains an unsafe deserialization vulnerability (CWE-502) in its checkpoint loading mechanism that allows remote code execution when a user loads a maliciously crafted checkpoint file. The vulnerability affects NVIDIA Megatron-LM installations and can lead to code execution, privilege escalation, information disclosure, and data tampering with a CVSS score of 7.8. The attack requires local access and low privileges but no user interaction once the malicious file is loaded.

NVIDIA Megatron-LM contains an insecure deserialization vulnerability (CWE-502) during model inferencing that allows remote code execution when a user loads a maliciously crafted input file. This vulnerability has a CVSS score of 7.8 and requires local access with low privileges but no user interaction, enabling attackers to execute arbitrary code, escalate privileges, disclose sensitive information, and tamper with data. The vulnerability affects NVIDIA's large language model training framework widely used in AI research and production environments.

NVIDIA Megatron-LM contains an unsafe deserialization vulnerability (CWE-502) in its checkpoint loading functionality that allows remote code execution when a user is tricked into loading a maliciously crafted checkpoint file. The vulnerability affects NVIDIA Megatron-LM installations and can lead to code execution, privilege escalation, information disclosure, and data tampering with a CVSS score of 7.8. There is no current indication of active exploitation in CISA's KEV catalog, and EPSS data was not provided in the intelligence sources.

IDrive's id_service.exe process, which runs with SYSTEM-level privileges on Windows systems, is vulnerable to privilege escalation and arbitrary code execution due to insecure file handling. The vulnerability affects IDrive Cloud Backup Client for Windows across multiple versions, as the elevated service process reads UTF16-LE encoded configuration files from C:\ProgramData\IDrive\ that are writable by standard unprivileged users. An attacker with local user access can modify these files to inject arbitrary executable paths, causing id_service.exe to execute malicious code with SYSTEM privileges, resulting in complete system compromise. While CVSS and EPSS scores are not available, this represents a critical local privilege escalation vector with trivial exploitability due to the file write permissions on the ProgramData directory.

A Remote Code Execution vulnerability exists in Craft CMS versions 4.x and 5.x that bypasses previous security patches for behavior injection attacks. An authenticated user with control panel access can exploit an unsanitized fieldLayouts parameter in the ElementIndexesController to inject malicious Yii2 behaviors and achieve arbitrary code execution. While no active exploitation (KEV) is documented, a patch is available and the vulnerability requires only low-privilege authenticated access, making it a significant risk for deployments with multiple control panel users.

A signed integer overflow vulnerability exists in the libtiff library's putcontig8bitYCbCr44tile function that leads to out-of-bounds heap writes through incorrect memory pointer calculations. Red Hat Enterprise Linux versions 6, 7, 8, 9, and 10 are confirmed affected. An attacker can exploit this by tricking a user into opening a specially crafted TIFF file, potentially achieving arbitrary code execution or causing application crashes.

NGINX Open Source and NGINX Plus contain a buffer over-read or over-write vulnerability in the ngx_http_mp4_module that can lead to NGINX worker process termination or potentially remote code execution. An attacker with local access and the ability to supply a specially crafted MP4 file for processing can exploit this flaw when the mp4 directive is enabled in the configuration. The vulnerability has a CVSS score of 7.8 with high impact on confidentiality, integrity, and availability, though exploitation requires local access (AV:L) and low-level privileges (PR:L).

An unauthenticated shell injection vulnerability exists in Langflow's GitHub Actions CI/CD workflows, allowing attackers to execute arbitrary commands by crafting malicious branch names or pull request titles. Langflow versions prior to 1.9.0 are affected, specifically the langflow-ai:langflow product. A proof-of-concept exploit exists demonstrating secret exfiltration via crafted branch names, enabling attackers to steal GITHUB_TOKEN credentials and potentially compromise the supply chain without any authentication required.

Multiple memory safety bugs affecting Firefox, Firefox ESR, and Thunderbird browsers present a critical remote code execution risk through memory corruption vulnerabilities. The affected versions include Firefox below 149, Firefox ESR below 115.34 and 140.9, Thunderbird ESR 140.8, Firefox 148, and Thunderbird 148. These memory safety issues demonstrate evidence of exploitable memory corruption that could allow attackers to execute arbitrary code on affected systems, though no public exploit or active KEV confirmation is currently documented.

Multiple memory safety bugs in Firefox 148 and Thunderbird 148 allow attackers to trigger memory corruption with potential for arbitrary code execution. Firefox versions prior to 149 are vulnerable, as confirmed by Mozilla security advisories. The vulnerability requires no user interaction beyond normal browsing and represents a critical elevation risk due to the presume-exploitable nature of the underlying memory corruption issues.

Multiple memory safety bugs affecting Firefox, Firefox ESR, Thunderbird, and Thunderbird ESR allow remote attackers to achieve arbitrary code execution through memory corruption vulnerabilities. Firefox versions prior to 149 and Firefox ESR versions prior to 140.9 are confirmed affected, with evidence suggesting these memory corruption issues could be exploited under sufficient effort. The vulnerability class encompasses buffer overflow and memory safety defects that demonstrate exploitation potential, though no active public exploitation has been documented at this time.

A code injection vulnerability exists in dendibakh perf-ninja's Lua modules (specifically in ldo.C within labs/misc/pgo), allowing improper control of code generation that can lead to remote code execution. The vulnerability affects all versions of perf-ninja as indicated by the CPE specification. An attacker can exploit this flaw to inject and execute arbitrary code, with a vendor patch now available to remediate the issue.

Remote code execution in Google Chrome's Federated Credential Management (FedCM) prior to version 146.0.7680.165 enables unauthenticated attackers to execute arbitrary code within the browser sandbox through a malicious HTML page. This use-after-free vulnerability in memory management affects Chrome on all supported platforms and requires only user interaction to trigger. A patch is available in Chrome 146.0.7680.165 and later.

Sandboxed code execution in Google Chrome's WebGPU implementation (prior to 146.0.7680.165) stems from a use-after-free memory vulnerability that can be triggered via malicious HTML pages. An unauthenticated remote attacker can exploit this to execute arbitrary code within the Chrome sandbox without user interaction beyond viewing a crafted webpage. A patch is available for affected users.

Remote code execution in llama.cpp prior to commit b7824 is possible through a crafted GGUF file that exploits an integer overflow in the `ggml_nbytes` function, causing heap buffer overflow during tensor processing. An attacker can bypass memory validation by specifying tensor dimensions that cause the size calculation to underflow dramatically, allowing memory corruption and potential code execution. The vulnerability affects Debian and other systems running vulnerable versions of llama.cpp, with no patch currently available.

The Jupiter X Core plugin for WordPress contains an unrestricted file upload vulnerability allowing authenticated users with Subscriber-level privileges or higher to upload dangerous file types including .phar, .svg, .dfxp, and .xhtml files. This stems from missing authorization checks in the import_popup_templates() function and insufficient file type validation in the upload_files() function. Successful exploitation leads to Remote Code Execution on Apache servers with mod_php configured to execute .phar files, or Stored Cross-Site Scripting attacks via malicious SVG and other file types on any server configuration.

The Woocommerce Custom Product Addons Pro plugin for WordPress contains a critical remote code execution vulnerability caused by unsafe use of PHP's eval() function when processing custom pricing formulas. All versions up to and including 5.4.1 are affected, allowing unauthenticated attackers to execute arbitrary PHP code on the server by submitting malicious input to WCPA text fields configured with custom pricing formulas. With a CVSS score of 9.8, this represents a maximum severity issue requiring immediate attention, though EPSS and KEV status data are not provided in the available intelligence.

A critical remote code execution vulnerability exists in PTC Windchill PDMLink and PTC FlexPLM products due to unsafe deserialization of untrusted data, allowing unauthenticated attackers to execute arbitrary code on affected systems. The vulnerability affects multiple versions of both products spanning from version 11.0 through 13.1.3.0 for Windchill and 11.0 through 13.0.3.0 for FlexPLM. An attacker can craft malicious serialized objects that, when deserialized by the vulnerable application, trigger code execution with the privileges of the Windchill or FlexPLM service account.

An authenticated code injection vulnerability exists in the Code Study Plugin component of OpenSource Workshop Connect-CMS that allows authenticated users to execute arbitrary code on the server. Both the 1.x series (versions up to 1.41.0) and 2.x series (versions up to 2.41.0) are affected. With a CVSS score of 8.8 (High severity), this vulnerability enables remote code execution and information disclosure with low attack complexity and no user interaction required.

An arbitrary file-write vulnerability exists in Pega Browser Extension (PBE) affecting Pega Robot Studio developers using versions 22.1 or R25 who automate Google Chrome and Microsoft Edge browsers. A threat actor can craft a malicious website that, when visited by a developer during interrogation mode in Robot Studio, executes arbitrary file-write operations on the developer's system. This vulnerability does not affect end-user Robot Runtime deployments, limiting its blast radius to development environments.

WWBN AVideo versions up to and including 26.0 contain a critical file upload vulnerability in the ImageGallery::saveFile() method that allows authenticated attackers to upload polyglot files (JPEG with embedded PHP code) and achieve Remote Code Execution. The vulnerability exploits a mismatch between MIME type validation (which checks file content) and filename extension handling (which trusts user input), allowing attackers to bypass security controls and execute arbitrary code on the server. A patch is available in commit 345a8d3ece0ad1e1b71a704c1579cbf885d8f3ae, and the issue has been publicly disclosed via GitHub Security Advisory GHSA-wxjw-phj6-g75w.

WWBN AVideo versions up to and including 26.0 contain an unauthenticated path traversal vulnerability in the locale API endpoint that allows arbitrary PHP file inclusion under the web root. Attackers can achieve confirmed file disclosure and code execution by including existing PHP files, with potential escalation to full remote code execution if they can upload or control PHP files elsewhere in the application tree. The vulnerability has a CVSS score of 8.6 and requires no authentication or user interaction to exploit, though no patch is currently available and there is no evidence of active exploitation in KEV data.

A critical command injection vulnerability exists in DigitalOcean Droplet Agent through version 1.3.2, where the troubleshooting actioner component processes metadata from the metadata service endpoint without adequate input validation, allowing attackers who can control metadata responses to inject and execute arbitrary OS commands with root privileges. An attacker can trigger the vulnerability by sending a TCP packet with specific sequence numbers to the SSH port, causing the agent to fetch and execute malicious commands from the metadata service, potentially leading to complete system compromise, data exfiltration, and lateral movement across cloud infrastructure. A public proof-of-concept exists at https://github.com/poxsky/CVE-2026-24516-DigitalOcean-RCE, indicating active research and potential exploitation risk.

The Performance Monitor plugin for WordPress contains a Server-Side Request Forgery (SSRF) vulnerability in its REST API endpoint that allows unauthenticated attackers to make arbitrary web requests to internal services using dangerous protocols including Gopher. Versions up to and including 1.0.6 are affected. This vulnerability can be chained with services like Redis to achieve Remote Code Execution, making it a critical security concern despite the 7.2 CVSS score.

OpenClaw versions prior to 2026.2.22 contain a shell environment variable injection vulnerability in the system.run function that allows attackers to bypass command allowlist protections. Authenticated remote attackers with low privileges can inject malicious shell startup files (.bash_profile, .zshenv) via unsanitized HOME and ZDOTDIR variables to achieve arbitrary code execution before allowlisted commands execute. A patch is available from the vendor via GitHub commit c2c7114ed39a547ab6276e1e933029b9530ee906.

Unsafe IOCTL handling in the DDK kernel module allows local attackers with limited privileges to bypass GPU memory protections and write to arbitrary physical memory through race condition exploitation. This privilege escalation vulnerability affects systems using the vulnerable DDK driver and requires no user interaction to trigger. No patch is currently available.

Remote code execution in Python allows authenticated users with SETTINGS permission to modify the reconnect.script configuration parameter without restriction, which is then passed unsanitized to subprocess.run() enabling arbitrary command execution. The vulnerability exists due to insufficient input validation in the set_config_value() API endpoint, which only restricts the general.storage_folder setting while leaving other security-critical options like reconnect.script unprotected. An attacker with non-admin SETTINGS privileges can exploit this to achieve full system compromise on the affected Python installation.

A Cross-Site Request Forgery (CSRF) vulnerability in the AVideo platform's plugin upload endpoint allows unauthenticated attackers to achieve Remote Code Execution by tricking authenticated administrators into visiting a malicious webpage. The vulnerability combines missing CSRF token validation on the pluginImport.json.php endpoint with explicitly configured SameSite=None session cookies over HTTPS, enabling cross-origin session hijacking. A proof-of-concept exploit has been published demonstrating full compromise by uploading a malicious plugin containing a PHP webshell.

The Kali Forms plugin for WordPress contains a critical remote code execution vulnerability that allows unauthenticated attackers to execute arbitrary code on the server. All versions up to and including 2.4.9 are affected, including the popular 'Kali Forms - Contact Form & Drag-and-Drop Builder' plugin by WPChill. The vulnerability carries a critical CVSS score of 9.8 due to its network-based attack vector, low complexity, and lack of required authentication or user interaction.