Remote Code Execution

Expression injection in Nuclei's template evaluation engine allows malicious HTTP servers to inject and execute DSL expressions via response data reused in multi-step templates. When the `-env-vars` flag is enabled (off by default), attackers can exfiltrate host environment variables including API keys and credentials; without this flag, injected expressions may trigger helper functions with limited security impact. Nuclei v3.8.0+ patches the vulnerability by collecting expressions from template source before placeholder substitution, preventing response-derived data from being reinterpreted as executable DSL syntax.

Unauthenticated attackers can execute remote code and read arbitrary files in Xerte Online Toolkits 3.15 and earlier via a missing authentication flaw in the elFinder connector endpoint. The vulnerability stems from a logic error where HTTP redirects for unauthenticated requests fail to terminate PHP execution, allowing full server-side processing of file operations. Attackers can create directories, upload files, rename, duplicate, overwrite, and delete files in project media directories without authentication. When chained with path traversal and extension blocklist bypasses, this enables complete system compromise. VulnCheck identified the flaw, and vendor patches are available via three GitHub commits addressing versions 3.13.0 through 3.15.0.

Path traversal in Xerte Online Toolkits' elFinder connector allows authenticated attackers to move files to arbitrary filesystem locations, enabling application file overwrites, stored XSS, or chained remote code execution. Affects versions 3.15 and earlier through unsanitized rename operations at /editor/elfinder/php/connector.php. Vendor patches available via GitHub commits 02661be, 507d55c, and 17e4f94. CVSS 7.1 with low attack complexity and low privileges required. No public exploitation confirmed (SSVC: exploitation=none), but attack is not automatable per CISA framework.

Remote code execution in ci4ms content management system allows authenticated backend users with theme creation permissions to write arbitrary PHP files via Zip Slip path traversal. A working proof-of-concept demonstrates uploading a malicious theme archive containing path-traversal entries (../../public/shell.php) that bypass extraction directory boundaries, placing executable code under the web root. Vendor-released patch available in version 0.31.5.0. No CISA KEV listing or EPSS data available, but publicly disclosed PoC significantly lowers exploitation barrier for attackers with valid credentials.

Remote code execution in ci4ms (CodeIgniter 4 Management System) versions prior to 0.31.5.0 allows authenticated backend users with backup creation permissions to write PHP webshells to the public web root via Zip Slip path traversal during backup restoration. The vulnerability is confirmed actively exploited (CISA KEV) with publicly available exploit code exists. CVSS 9.4 (Critical) aligns with the real-world risk, as exploitation requires only low-privilege authentication and the affected route is exempt from CSRF protection, enabling drive-by attacks against logged-in administrators. Vendor-released patch version 0.31.5.0 addresses the flaw by implementing path validation during ZIP extraction.

Privilege escalation to root in uutils coreutils chroot utility allows low-privileged local attackers with write access to the chroot target directory to execute arbitrary code via malicious NSS module injection. The vulnerability triggers when --userspec option causes getpwnam() to load attacker-controlled shared libraries from the new root before dropping privileges, enabling container escape or full system compromise on glibc-based systems. CVSS 7.8 with Scope Changed indicates host compromise from containerized environments. SSVC framework confirms POC availability and total technical impact, though exploitation requires specific configuration (writable NEWROOT) and is not automatable.

ELBA5 5.8.0 contains a remote code execution vulnerability that allows attackers to obtain database credentials and execute arbitrary commands with SYSTEM level permissions. Rated critical severity (CVSS 9.3), this vulnerability is remotely exploitable, no authentication required, low attack complexity. Public exploit code available and no vendor patch available.

ThinkPHP 5.0.23 contains a remote code execution vulnerability that allows unauthenticated attackers to execute arbitrary PHP code by invoking functions through the routing parameter. Rated critical severity (CVSS 9.3), this vulnerability is remotely exploitable, no authentication required, low attack complexity. Public exploit code available and no vendor patch available.

LanSpy 2.0.1.159 contains a local buffer overflow vulnerability that allows attackers to overwrite the instruction pointer by supplying oversized input to the scan field. Rated high severity (CVSS 8.6), this vulnerability is no authentication required, low attack complexity. Public exploit code available and no vendor patch available.

LanSpy 2.0.1.159 contains a local buffer overflow vulnerability in the scan section that allows local attackers to execute arbitrary code by exploiting structured exception handling mechanisms. Rated high severity (CVSS 8.6), this vulnerability is no authentication required, low attack complexity. Public exploit code available and no vendor patch available.

Iperius Backup 5.8.1 contains a local buffer overflow vulnerability in the structured exception handling (SEH) mechanism that allows local attackers to execute arbitrary code by supplying a malicious. Rated high severity (CVSS 8.6), this vulnerability is no authentication required, low attack complexity. Public exploit code available and no vendor patch available.

MAGIX Music Editor 3.1 contains a buffer overflow vulnerability in the FreeDB Proxy Options dialog that allows local attackers to execute arbitrary code by exploiting structured exception handling. Rated high severity (CVSS 8.6), this vulnerability is no authentication required, low attack complexity. Public exploit code available and no vendor patch available.

Terminal Services Manager 3.1 contains a stack-based buffer overflow vulnerability in the computer names field that allows local attackers to execute arbitrary code by triggering structured exception. Rated high severity (CVSS 8.6), this vulnerability is no authentication required, low attack complexity. Public exploit code available and no vendor patch available.

Path traversal vulnerability in Poetry's tar extraction function allows arbitrary file writes when processing untrusted source distributions on Python 3.10.0-3.10.12 and 3.11.0-3.11.4, where the tarfile.data_filter safety mechanism is absent or broken. The vulnerability is triggered during dependency resolution (poetry add --lock) or installation before the build backend executes, enabling attackers to write files outside the intended extraction directory via crafted tar member paths, symlinks, or hardlinks in malicious sdists.

Remote attackers can corrupt PowerDNS Authoritative Server configuration via specially crafted DNS NOTIFY requests, causing persistent denial of service requiring manual administrator intervention. The attack adds malformed secondary domains to the bind backend, rendering the configuration invalid and preventing the server from restarting. No active exploitation confirmed at time of analysis, but the network-accessible attack vector and lack of authentication requirements elevate risk for internet-facing authoritative DNS servers.

Remote code execution in InstructLab affects Red Hat Enterprise Linux AI 3 when users download or train models from HuggingFace Hub. The linux_train.py script hardcodes trust_remote_code=True, allowing attackers to execute arbitrary Python code by hosting malicious models on HuggingFace and convincing users to run ilab train, download, or generate commands. This configuration weakness enables complete system compromise through social engineering attacks. CVSS 8.8 with network vector but requires user interaction, reducing automatic exploitation risk. No active exploitation (CISA KEV) or public POC identified at time of analysis.

Remote code execution in Red Hat Apache Camel Infinispan component allows low-privileged attackers to execute arbitrary code via unsafe deserialization in ProtoStream remote aggregation repository. Exploiting this vulnerability requires network access and low-privilege credentials but grants full system compromise affecting confidentiality, integrity, and availability. The attack complexity is rated high (AC:H), suggesting specific configuration or timing requirements. No active exploitation confirmed at time of analysis (not in CISA KEV), and public exploit code status is unknown.

Heap buffer overflow in GNU Binutils XCOFF linker allows arbitrary code execution when a local user processes a malicious object file. Red Hat Enterprise Linux versions 6 through 10 are confirmed affected via CPE data. CVSS 7.8 reflects local attack vector requiring user interaction (opening/linking the crafted file). No active exploitation confirmed (not in CISA KEV), and no public proof-of-concept identified at time of analysis. Real-world risk depends heavily on whether development workflows involve linking untrusted XCOFF files, which is uncommon outside AIX/PowerPC cross-compilation scenarios.

Arbitrary file write in HTTP Headers plugin for WordPress versions ≤1.19.2 enables authenticated administrators to achieve remote code execution by manipulating htpasswd file path configuration and injecting PHP code via unsanitized username input. Administrators can set a malicious file path (e.g., webroot/shell.php) through 'hh_htpasswd_path' option and inject executable code via the 'hh_www_authenticate_user' field, which is written directly to disk without validation. Wordfence disclosure includes direct source code references showing the vulnerable apache_auth_credentials() and update_auth_credentials() functions. No public exploit code or active exploitation confirmed at time of analysis.

Remote code execution in Progress Telerik UI for ASP.NET AJAX via insecure deserialization in the RadFilter control allows unauthenticated remote attackers to execute arbitrary code on the server by tampering with exposed client-side filter state. Affected versions span 2024.4.1114 through 2026.1.421. EPSS data not available; no public exploit or CISA KEV listing identified at time of analysis. The CVSS 8.1 (High) reflects network accessibility but 'High' attack complexity (AC:H), indicating successful exploitation requires specific conditions beyond simple network access.

F Prime framework before version 4.2.0 allows remote code execution via integer overflow in bounds checking combined with path traversal in file upload functionality. An attacker sending a crafted DataPacket with byteOffset=0xFFFFFF9C and dataSize=100 causes U32 addition to wrap to zero, bypassing the fileSize validation. This enables writing arbitrary data to any file at any offset on the target system, leading to remote code execution on embedded spaceflight and other critical systems. CISA KEV status and active exploitation unknown; vendor patch available in version 4.2.0.

Remote code execution in AVideo versions 29.0 and below allows unauthenticated attackers to execute arbitrary shell commands on the server via command injection in the CloneSite plugin's cloneServer.json.php endpoint. Attackers exploit unsanitized user input in the 'url' parameter that gets directly concatenated into a wget command executed through PHP's exec() function. With CVSS 8.9 (AV:N/AC:L/PR:N/UI:N) and proof-of-concept exploitation confirmed (E:P), this represents a critical risk requiring immediate patching. Fix available in commit 473c609fc2defdea8b937b00e86ce88eba1f15bb.

Tekton Pipelines project provides k8s-style resources for declaring CI/CD-style pipelines. From 1.0.0 to before 1.11.0, the git resolver's revision parameter is passed directly as a positional argument to git fetch without any validation that it does not begin with a - character. Because git parses flags from mixed positional arguments, an attacker can inject arbitrary git fetch flags such as --upload-pack=<binary>. Combined with the validateRepoURL function explicitly permitting URLs that begin with / (local filesystem paths), a tenant who can submit ResolutionRequest objects can chain these two behaviors to execute an arbitrary binary on the resolver pod. The tekton-pipelines-resolvers ServiceAccount holds cluster-wide get/list/watch on all Secrets, so code execution on the resolver pod enables full cluster-wide secret exfiltration. This vulnerability is fixed in 1.11.1.

## Abstract Trend Micro's Zero Day Initiative has identified a vulnerability affecting FlowiseAI Flowise. ## Vulnerability Details - **Version tested:** 3.0.13 - **Installer file:** https://github.com/FlowiseAI/Flowise - **Platform tested:** Ubuntu 25.10 ## Analysis This vulnerability allows remote attackers to execute arbitrary code on affected installations of FlowiseAI Flowise. Authentication is not required to exploit this vulnerability. The specific flaw exists within the `run` method of the `CSV_Agents` class. The issue results from the lack of proper sandboxing when evaluating an LLM-generated Python script. An attacker can leverage this vulnerability to execute code in the context of the user running the server. ### Product Information FlowiseAI Flowise version 3.0.13 - https://github.com/FlowiseAI/Flowise ### Setup Instructions ```bash npm install -g flowise@3.0.13 npx flowise start ``` ### Root Cause Analysis FlowiseAI Flowise is an open source low-code tool for developers to build customized large language model (LLM) applications and AI agents. It supports integration with various LLMs, data sources, and tools in order to facilitate rapid development and deployment of AI solutions. Flowise offers a web interface with a drag-and-drop editor, as well as an API, through an Express web server accessible over HTTP on port 3000/TCP. One such feature of Flowise is the ability to create chatflows. Chatflows use a drag-and-drop editor that allows a developer to place nodes which control how an interaction with an LLM will occur. One such node is the CSV Agent node that represents an Agent used to answer queries on a provided CSV file. When a user makes a query against a chatflow using the CSV Agent node, the `run` method of the `CSV_Agents` class is called. This method first reads the contents of the CSV file passed to the node and converts it to a base64 string. It then sets up a pyodide environment and creates a Python script to be executed in this environment. This Python script uses pandas to extract the column names and their types from the provided CSV file. The method then creates a system prompt for an LLM using this data as follows: ``` You are working with a pandas dataframe in Python. The name of the dataframe is df. The columns and data types of a dataframe are given below as a Python dictionary with keys showing column names and values showing the data types. {dict} I will ask question, and you will output the Python code using pandas dataframe to answer my question. Do not provide any explanations. Do not respond with anything except the output of the code. Security: Output ONLY pandas/numpy operations on the dataframe (df). Do not use import, exec, eval, open, os, subprocess, or any other system or file operations. The code will be validated and rejected if it contains such constructs. Question: {question} Output Code: ``` Where `{dict}` is the extracted column names and `{question}` is the initial prompt provided by the user. This system prompt is sent to an LLM in order for it to generate a Python script based on the user's prompt, and the LLM-generated response is stored in a variable named `pythonCode`. The method then evaluates the `pythonCode` variable in a pyodide environment. While the LLM-generated Python script is evaluated in a non-sandboxed environment, there is a list of forbidden patterns that are checked before the script is executed on the server. The function `validatePythonCodeForDataFrame()` enumerates through a list named `FORBIDDEN_PATTERNS`, which contains pairs of regex patterns and reasons. Each regex pattern is run against the Python script, and if the pattern is found in the script, the script is invalidated and is not run, responding to the request with a reason for rejection. The input validation can be bypassed, which can still lead to running arbitrary OS commands on the server. An example of this is the pattern `/\bimport\s+(?!pandas|numpy\b)/g`, which intends to search for lines of code that import a module other than pandas or numpy. This can be bypassed by importing along with pandas or numpy. For example, consider the following lines of code: ```python import pandas as np, os as pandas pandas.system("xcalc") ``` Here, pandas is imported, but so is the `os` module, with `pandas` as its alias. OS commands can then be invoked with `pandas.system()`. Using prompt injection techniques, an unauthenticated attacker with the ability to send prompts to a chatflow using the CSV Agent node may convince an LLM to respond with a malicious Python script that executes attacker-controlled commands on the Flowise server. It is also possible for an authenticated attacker to exploit this vulnerability by specifying an attacker-controlled server in a chatflow. This server would respond to prompts with an attacker-controlled Python script instead of an LLM-generated response, which would then be evaluated on the server. ### Relevant Source Code #### `packages/components/nodes/agents/CSVAgent/core.ts` ```ts import type { PyodideInterface } from 'pyodide' import * as path from 'path' import { getUserHome } from '../../../src/utils' let pyodideInstance: PyodideInterface | undefined export async function LoadPyodide(): Promise<PyodideInterface> { if (pyodideInstance === undefined) { const { loadPyodide } = await import('pyodide') const obj: any = { packageCacheDir: path.join(getUserHome(), '.flowise', 'pyodideCacheDir') } pyodideInstance = await loadPyodide(obj) await pyodideInstance.loadPackage(['pandas', 'numpy']) } return pyodideInstance } export const systemPrompt = `You are working with a pandas dataframe in Python. The name of the dataframe is df. The columns and data types of a dataframe are given below as a Python`*

WWBN AVideo is an open source video platform. In versions 29.0 and prior, the YPTSocket plugin's WebSocket server relays attacker-supplied JSON message bodies to every connected client without sanitizing the `msg` or `callback` fields. On the client side, `plugin/YPTSocket/script.js` contains two `eval()` sinks fed directly by those relayed fields (`json.msg.autoEvalCodeOnHTML` at line 568 and `json.callback` at line 95). Because tokens are minted for anonymous visitors and never revalidated beyond decryption, an unauthenticated attacker can broadcast arbitrary JavaScript that executes in the origin of every currently-connected user (including administrators), resulting in universal account takeover, session theft, and privileged action execution. Commit c08694bf6264eb4decceb78c711baee2609b4efd contains a fix.

Path traversal in WWBN AVideo 29.0 and earlier allows authenticated administrators (or CSRF-tricked admins) to write arbitrary PHP files anywhere on the server filesystem, achieving remote code execution. The locale save endpoint fails to sanitize the 'flag' parameter used in file path construction and lacks CSRF protection despite SameSite=None cookies, enabling straightforward exploitation by lower-privilege attackers who chain CSRF against admin sessions. Upstream fix committed to GitHub (57f89ffb) but released patched version not independently confirmed. CVSS 8.7 reflects high impact but requires privileged access - real-world risk depends heavily on admin session hijacking opportunities.

Remote code execution in Atlassian Bamboo Data Center versions 9.6.0 through 12.1.0 allows authenticated attackers to execute arbitrary OS commands via command injection vulnerability. The attack requires low-privilege authentication (PR:L) but no user interaction, enabling complete system compromise across confidentiality, integrity, and availability with cross-scope impact (SC:H/SI:H/SA:H indicating container escape or lateral movement potential). Atlassian has released patches for three major version branches (9.6.25, 10.2.18, 12.1.6). No active exploitation confirmed in CISA KEV at time of analysis, though the authenticated nature and critical CVSS 9.4 score warrant immediate patching for internet-exposed instances with broad user access.

Local privilege escalation in CivetWeb v1.16 service allows authenticated users to execute arbitrary code with SYSTEM privileges via unquoted service path exploitation. The Windows service configuration lacks quotes around 'C:\Program Files\CivetWeb\CivetWeb.exe', enabling attackers to place malicious executables in directories scanned before the intended path (e.g., 'C:\Program.exe' or 'C:\Program Files\CivetWeb.exe'). No public exploit identified at time of analysis, though EPSS data not available. Patch available per vendor advisory from INCIBE.

Multiple memory corruption bugs in Firefox ESR 140.9, Firefox 149, Thunderbird ESR 140.9, and Thunderbird 149 could enable remote code execution against users visiting malicious websites. Mozilla has fixed these memory safety vulnerabilities in Firefox 150 and Firefox ESR 140.10, with vendor advisories (MFSA2026-30, MFSA2026-32, MFSA2026-33, MFSA2026-34) confirming patches are available. EPSS score of 0.05% (14th percentile) indicates low observed exploitation probability, and no public exploit identified at time of analysis, though SSVC framework assesses total technical impact if successfully weaponized.

Use-after-free memory corruption (CWE-416) in Mozilla Firefox 149 and ESR 115.34/140.9, plus the shared Gecko engine in Thunderbird 149 and Thunderbird ESR 140.9, can lead to arbitrary code execution within the browser process when a victim renders attacker-controlled web content. This is a rolled-up batch of memory-safety bugs reported by Mozilla's own developers; Mozilla states some showed evidence of memory corruption presumed exploitable for code execution. There is no public exploit identified at time of analysis, the bug is not in CISA KEV, and EPSS is very low (0.06%, 17th percentile), consistent with the CVSS 7.5 rating being held down by high attack complexity (AC:H) and required user interaction (UI:R).

Memory corruption in Firefox 149 and Thunderbird 149 enables remote code execution when users interact with malicious web content. Mozilla patched 55 distinct memory safety bugs in Firefox 150, some demonstrating memory corruption that could be weaponized for arbitrary code execution. While no public exploit is confirmed, the CVSS score of 7.5 reflects high complexity requiring user interaction, with SSVC assessment indicating total technical impact despite no current automation or active exploitation.

Remote code execution in Quantum Networks router QN-I-470 allows authenticated attackers to execute arbitrary OS commands as root via command injection in the management CLI interface. The vulnerability stems from inadequate input sanitization, enabling low-privileged authenticated users to escalate privileges to root level. CVSS 8.7 (Critical) reflects network-accessible exploitation with low complexity, requiring only low-privilege authentication. No active exploitation (CISA KEV) or public exploit code identified at time of analysis, but the authenticated nature and CLI access requirement limits exploitation to users with existing device credentials.

Remote code execution with root privileges in Quantum Networks router QN-I-470 version 6.1.1.B1 allows adjacent network attackers to execute arbitrary OS commands through the management CLI interface via command injection. The vulnerability requires no authentication (CVSS PR:N) and exploits inadequate input sanitization (CWE-78). Adjacent network access (AV:A) limits attack surface to local network segments. No active exploitation (CISA KEV) or public exploit code identified at time of analysis, though EPSS data unavailable to assess real-world exploitation probability.

Command injection in Lawnchair's GitHub Actions workflow allows authenticated repository contributors to execute arbitrary code on GitHub-hosted CI/CD runners. The vulnerability affects Lawnchair for Android versions prior to commit fcba413f5 and stems from unsanitized workflow_dispatch inputs in release_update.yml. Authenticated attackers with repository write access can inject shell commands through workflow parameters, achieving full code execution in the build environment. A patch is available (commit fcba413f5), and the CVSS vector indicates this is a network-accessible, low-complexity attack requiring low privileges. CVSS v4.0 score of 7.4 reflects high confidentiality, integrity, and availability impact scoped to the vulnerable CI/CD system. EPSS data not provided; no CISA KEV listing at time of analysis.

Sandbox escape in Claude Code versions prior to 2.1.64 enables arbitrary file writes outside the workspace by exploiting symlink handling between sandboxed and unsandboxed processes, potentially leading to code execution. The vulnerability requires prompt injection to trigger malicious sandboxed code execution, creating an exploitable chain where neither component can independently breach the sandbox but their interaction does. EPSS score of 0.08% (23rd percentile) suggests limited real-world exploitation likelihood, and CISA SSVC indicates no known exploitation with non-automatable attack requirements. Version 2.1.64 patches this issue, auto-deployed to standard installations.

Malicious workspace plugins in OpenClaw versions before 2026.4.2 achieve arbitrary code execution by shadowing built-in channel IDs during workspace clone and setup operations. The vulnerability exploits a trust boundary flaw (CWE-829) where untrusted plugins execute before explicit user trust confirmation, requiring only that a victim clone a poisoned workspace repository. With CVSS 8.5 (High) and local attack vector requiring user interaction, real-world risk is moderate: EPSS probability sits at 0.01% (2nd percentile) with no confirmed active exploitation (not in CISA KEV), and SSVC assessment classifies it as non-automatable with total technical impact but no current exploitation.

Remote code execution in Dolibarr ERP 22.0.4 and earlier allows authenticated users with PHP content editing permissions to execute arbitrary OS commands on the server. The vulnerability stems from a bypassable blacklist-based filter for dangerous PHP functions in the Website module. Attack complexity is low (CVSS AV:N/AC:L/PR:L), requiring only valid low-privilege credentials. Public proof-of-concept code exists on GitHub, though CISA has not confirmed active exploitation. EPSS data is unavailable, but SSVC assessment indicates total technical impact with no current exploitation evidence.

Remote code execution in Visitor Management System 1.0 allows authenticated administrators to upload PHP webshells via two unvalidated file upload endpoints (admin_user_insert.php and update_1.php). The move_uploaded_file() function lacks MIME type, extension, and content validation, enabling direct server compromise. Public proof-of-concept exists (SSVC exploitation: POC). EPSS data not available, but the combination of network-accessible attack vector (AV:N) and total technical impact (SSVC) against a specific niche product suggests targeted exploitation risk rather than widespread automated attacks.

Authenticated users with restricted HTML/JavaScript editing permissions in Dolibarr ERP & CRM 22.0.4 and earlier can escalate privileges to execute arbitrary PHP code via the Website module. The vulnerability exploits inconsistent permission enforcement across input parameters during website page creation, allowing low-privileged authenticated users to bypass intended restrictions and inject PHP code. Public proof-of-concept exists on GitHub (PhDg1410), though no active exploitation is confirmed by CISA KEV. EPSS data unavailable, but the CVSS 8.8 score reflects high impact across confidentiality, integrity, and availability when exploited by authenticated insiders or compromised accounts.

Remote code execution in Spinnaker's Echo service (all versions prior to 2026.1.0, 2026.0.1, 2025.4.2, and 2025.3.2) allows authenticated attackers with low privileges to execute arbitrary system commands and access files through unrestricted Spring Expression Language (SPeL) injection in artifact processing. Unlike Spinnaker's Orca service which implemented SPeL sandbox restrictions, Echo permits full JVM class access, enabling attackers to invoke arbitrary Java classes for deep system compromise. The CVSS 9.9 score reflects network attack vector with low complexity, scope change to impact other components, and complete CIA triad compromise. EPSS and KEV data not available - exploitation status unknown but patches are available from Spinnaker project.

Remote code execution in Vvveb CMS 1.0.8 allows authenticated attackers with low privileges to upload PHP webshells disguised with .phtml extensions, bypassing file type restrictions to achieve full server compromise. The vulnerability stems from inadequate file upload validation in the media handler, enabling malicious files in publicly accessible directories. Upstream fix available via GitHub commit; EPSS data unavailable, no CISA KEV listing at time of analysis.

Remote code execution in Vvveb CMS v1.0.8 allows authenticated administrators to execute arbitrary system commands as www-data via a two-stage file upload attack. Attackers exploit a logic flaw in the media management file rename handler that fails to block .php and .htaccess extensions, enabling MIME type manipulation followed by PHP code execution. VulnCheck published an advisory and GitHub commit 6fb8eaa confirms upstream fix. No EPSS data available; no active exploitation confirmed at time of analysis.

Arbitrary file deletion in wpForo Forum plugin versions ≤3.0.5 allows authenticated attackers with subscriber-level privileges to delete critical WordPress files including wp-config.php, enabling remote code execution. The vulnerability chains two flaws: unvalidated file paths in custom profile fields and insufficient path sanitization before file deletion. Exploitation requires the wpForo User Custom Fields addon with at least one file-type custom field configured. CVSS 8.1 (High) with network attack vector, low complexity, and low privilege requirements. EPSS data and active exploitation status not available in current intelligence.

Remote code execution in OpenMage Magento LTS versions before 20.17.0 allows authenticated attackers to upload executable PHP files through product custom options by bypassing an incomplete file extension blocklist. The vulnerability exists because the upload filter only blocks `.php` and `.exe` extensions, permitting alternative PHP-executable extensions like `.phtml`, `.phar`, `.php3-.php7`, and `.pht`. Uploaded files land in the publicly accessible `media/custom_options/quote/` directory, enabling code execution on servers without explicit script execution restrictions. No active exploitation confirmed (not in CISA KEV), but public disclosure via GitHub Security Advisory increases exploitation likelihood. EPSS data not provided.

Remote code execution in OpenMage Magento LTS versions prior to 20.17.0 allows unauthenticated attackers to execute arbitrary code by uploading malicious phar archives disguised as images and triggering PHP deserialization via phar:// stream wrappers. The attack requires high complexity (AC:H) to exploit successfully. EPSS data not available, but exploitation requires specific conditions around file upload and path manipulation. Vendor patch available in version 20.17.0, confirmed by GitHub security advisory GHSA-fg79-cr9c-7369.

Remote code execution in Vvveb CMS versions prior to 1.0.8.1 allows unauthenticated attackers to inject arbitrary PHP code through the installation endpoint's subdir parameter, which is written directly into env.php without sanitization. The vulnerability enables complete system compromise as the web server user with no authentication required. Publicly available patch exists (version 1.0.8.1) with detailed fix commit reference.

Privilege escalation in Vvveb CMS versions prior to 1.0.8.1 allows authenticated low-privileged users to inject role_id=1 into profile save requests, escalating to Super Administrator and enabling plugin upload for remote code execution. Vendor patch available in version 1.0.8.1. CVSS 8.8 reflects high impact across confidentiality, integrity, and availability. EPSS data not provided; KEV status unknown. Public disclosure via VulnCheck advisory with commit-level fix details increases likelihood of exploitation attempts.

Stored cross-site scripting in Vvveb prior to 1.0.8.1 allows authenticated users with media upload and rename permissions to execute arbitrary JavaScript in administrator browsers by bypassing MIME type validation with a GIF89a header prepend, renaming files to .html extensions, and injecting malicious payloads that can create backdoor accounts or upload remote code execution plugins. Publicly available exploit code exists and vendor-released patch 1.0.8.1 is available. Real-world risk is moderate due to authentication requirement and required user interaction (administrator must visit malicious page), but privilege escalation path to RCE via plugin upload makes this a critical persistence vector.

Remote code execution in SGLang 0.5.9's /v1/rerank endpoint allows unauthenticated attackers to execute arbitrary code by loading specially crafted model files with malicious Jinja2 templates. The vulnerability stems from unsandboxed rendering of tokenizer.chat_template fields, enabling template injection attacks. Publicly available exploit code exists (GitHub POC by Stuub). With CVSS 9.8 (AV:N/AC:L/PR:N/UI:N) and SSVC ratings of automatable with total technical impact, this represents critical risk for exposed SGLang deployments handling untrusted model files.

OS command injection in Progress LoadMaster and related ADC products allows authenticated administrators with 'All' permissions to execute arbitrary commands via malicious WAF rule file uploads. The attacker exploits unsanitized input during the file upload process in the web UI. With CVSS 8.4 and scope change to 'Changed', successful exploitation enables complete system compromise beyond the vulnerable component. No active exploitation confirmed (not in CISA KEV) and no public POC identified at time of analysis. EPSS data not available for risk assessment.

Command injection in Progress LoadMaster, ECS Connections Manager, Object Scale Connection Manager, and MOVEit WAF allows authenticated administrators with 'VS Administration' privileges to execute arbitrary operating system commands on the appliance via unsanitized input to the 'aclcontrol' API command. CVSS 8.4 reflects high-privilege requirement but scope change indicates container escape or cross-boundary impact. EPSS data not provided. No public exploit identified at time of analysis. Vendor-released patch: version 7.2.63.0 for all affected products per Progress advisory.

OS command injection in Progress LoadMaster, MOVEit WAF, ECS Connections Manager, and Object Scale Connection Manager API allows authenticated attackers with 'All' permissions to execute arbitrary commands on appliances via unsanitized input in the 'killsession' API endpoint. CVSS 8.4 (High) reflects adjacent network access vector and high privileges requirement, limiting exploitation to administrators or compromised admin accounts. CISA SSVC assessment indicates no active exploitation, non-automatable attack, but total technical impact. EPSS data not provided, but privilege requirements significantly reduce real-world attack surface compared to unauthenticated RCE vulnerabilities.

Command injection in Progress LoadMaster and related ADC products allows authenticated attackers with Geo Administration permissions to execute arbitrary OS commands on appliances via the unsanitized 'addcountry' API parameter. Affects LoadMaster, ECS Connections Manager, MOVEit WAF, and Object Scale Connection Manager versions prior to 7.2.63.0. EPSS data unavailable; not listed in CISA KEV. CVSS 8.4 reflects high impact (complete system compromise) but requires adjacent network access and high-privilege authentication, significantly constraining real-world exploitation scenarios. Vendor has released patches addressing all affected products.

Local privilege escalation in SKYSEA Client View (≤21.200.07j) and SKYMEC IT Manager (≤2024.005.10a) allows low-privileged users to execute arbitrary code with administrative privileges by exploiting insecure installation folder permissions. Attackers can write malicious files into the product directory, achieving full system compromise. EPSS score of 0.01% (2nd percentile) indicates low likelihood of widespread exploitation despite CVSS 8.5 severity. No active exploitation confirmed; CISA SSVC assessment marks exploitation status as 'none' and automatable as 'no', suggesting targeted attack potential rather than mass exploitation risk.

DSL expression injection in ProjectDiscovery Nuclei before 3.8.0 allows remote code execution when using the -env-vars flag with multi-step templates against untrusted targets. An attacker can inject malicious expressions into environment variables that are evaluated as Nuclei DSL code, achieving arbitrary code execution with the privileges of the Nuclei process. This vulnerability requires non-default configuration (explicit -env-vars usage) and high attack complexity, limiting real-world impact despite the RCE tag.

Remote code execution in ASUSTOR ADM (ASUSTOR Data Master) operating system versions 4.1.0-4.3.3.RR42 and 5.0.0-5.1.2.REO1 allows authenticated administrators to inject arbitrary OS commands via the PPTP VPN Clients web interface. The command injection (CWE-78) bypasses the restricted web environment, enabling full system compromise. Attack complexity is low (AC:L) with network attack vector (AV:N), and CVSS 9.4 reflects critical impact across confidentiality, integrity, and availability. No active exploitation or public POC confirmed at time of analysis, though EPSS probability data not available.

Remote code execution in ASUSTOR ADM (4.1.0-4.3.3.RR42 and 5.0.0-5.1.2.REO1) allows authenticated high-privilege attackers to execute arbitrary code via stack-based buffer overflow in VPN client components. The vulnerability combines unbounded sscanf() calls with format string weaknesses (printf with user-controlled data), exploitable due to absent PIE and stack canary protections. EPSS exploitation probability is low (0.23%, 46th percentile) with no public exploit code identified at time of analysis, suggesting limited real-world targeting despite high CVSS score.

Remote code execution in ModelScope AgentScope up to version 1.0.18 allows unauthenticated network attackers to inject and execute arbitrary Python code or shell commands through the execute_python_code and execute_shell_command functions in src/AgentScope/tool/_coding/_python.py. Publicly available exploit code exists, and the vendor has not responded to early disclosure notifications, leaving all versions up to 1.0.18 unpatched and actively exploitable.

Remote code execution in silex technology SD-330AC and AMC Manager allows unauthenticated network attackers to execute arbitrary code via heap-based buffer overflow when processing redirect URLs. CVSS 9.3 critical severity with attack vector AV:N/AC:L/PR:N/UI:N indicates trivial exploitation against internet-facing devices. No public exploit identified at time of analysis, though JPCERT coordination suggests vendor-confirmed vulnerability. EPSS data not available; real-world risk depends on internet exposure of affected silex wireless bridge and management software installations.

Stack-based buffer overflow in silex technology's SD-330AC (Ver.1.42 and earlier) and AMC Manager (Ver.5.0.2 and earlier) enables authenticated remote attackers to execute arbitrary code on the device via maliciously crafted redirect URLs. Reported by JPCERT with vendor advisories published, though EPSS score of 0.04% (12th percentile) indicates low observed exploitation probability. No active exploitation confirmed (not in CISA KEV), and SSVC assessment marks exploitation status as 'none' despite the critical nature of remote code execution capability.

Local privilege escalation in DeepCool DeepCreative software version 1.2.7 and earlier allows unauthenticated attackers to execute arbitrary code with elevated privileges through malicious file processing. The vulnerability stems from insecure permission configuration (CWE-277) requiring user interaction to open a crafted file. Public exploit research exists on GitHub (uncle-hash repository), though CISA has not confirmed active exploitation. CVSS 7.8 indicates high severity, but EPSS data unavailable; SSVC framework rates technical impact as total with no confirmed exploitation and non-automatable attack path.

Code injection vulnerability in protobufjs (JavaScript protobuf library) allows authenticated attackers to execute arbitrary JavaScript code during protobuf object decoding by injecting malicious payloads into 'type' fields of protobuf definitions. Affects all versions before 7.5.5 and 8.0.1. CVSS 9.4 (Critical) reflects chained impact across multiple security boundaries (VC:H/VI:H/VA:H/SC:H/SI:H/SA:H), though exploitation requires authenticated access (PR:L) to inject malicious protobuf definitions. No active exploitation confirmed (not in CISA KEV); vendor-released patches available.

Deserialization vulnerability in Apache Airflow webserver (all versions before 3.2.0) allows network-accessible attackers to execute arbitrary code by injecting malicious XCom payloads, despite vendor-assigned Low severity due to the trusted Dag Author threat model. CVSS 9.8 Critical rating reflects unauthenticated network-based RCE capability (AV:N/PR:N), contradicting the description's trust assumption. EPSS 0.07% (22nd percentile) suggests low immediate exploitation likelihood. No active exploitation confirmed; vendor patch available in version 3.2.0 with public GitHub PR.

Remote code execution in iTerm2 through version 3.6.9 allows local attackers to execute arbitrary code by displaying a specially crafted text file when a malicious file with a conductor-protocol-compatible name exists in the working directory. The vulnerability exploits iTerm2's acceptance of SSH conductor protocol sequences (DCS 2000p and OSC 135) from terminal output without validating the source, enabling in-band signaling abuse where filenames themselves become attack vectors. CVSS 6.9 reflects local attack vector and high complexity, but practical exploitation requires user interaction (opening a file) combined with directory-resident malware.

Remote code execution in CMP - Coming Soon & Maintenance Plugin by NiteoThemes for WordPress (versions ≤4.1.16) allows authenticated attackers with Administrator-level privileges to upload and execute arbitrary PHP code via a malicious ZIP file. The vulnerability stems from insufficient capability checking (publish_pages instead of manage_options) and absent file validation in the cmp_theme_update_install AJAX action. CVSS 8.8 reflects high impact across confidentiality, integrity, and availability. No CISA KEV listing or public exploit code identified at time of analysis, suggesting limited real-world exploitation despite the high severity rating. Wordfence Threat Intelligence disclosed this vulnerability with detailed source code references.

gdown prior to version 5.2.2 allows remote attackers to write arbitrary files outside the intended extraction directory via maliciously crafted ZIP or TAR archives due to insufficient path traversal validation in the extractall functionality. An attacker can craft a malicious archive with path traversal sequences (e.g., ../ entries) in filenames, which when extracted by a user, permits file overwrite and potential remote code execution. The vulnerability requires user interaction (UI:R) to trigger extraction but affects all unauthenticated remote users downloading via the gdown library.

Path traversal in SecureDrop Client 0.17.4 and below allows a compromised SecureDrop Server to execute arbitrary code on journalist workstations by injecting malicious filenames during gzip archive extraction. The vulnerability enables overwriting the SQLite database and other critical files in the sd-app VM, potentially exposing decrypted source submissions. Fixed in version 0.17.5. No active exploitation confirmed (not in CISA KEV). CVSS 7.5 reflects high impact but complex attack chain requiring prior server compromise and user interaction. EPSS data not available, but real-world risk is constrained by the requirement to first breach a Tor-hidden, hardened server infrastructure.

Remote code execution in ChurchCRM <7.2.0 allows authenticated administrators to upload PHP webshells through the database backup restore function, which copies files from archive Images/ directories to web-accessible paths without extension filtering. The vulnerability includes a CSRF bypass enabling forced exploitation through cross-site request forgery. Exploitation requires high privileges (administrator account) but is network-accessible with low complexity (CVSS:3.1/AV:N/AC:L/PR:H). No active exploitation confirmed (not in CISA KEV). Vendor-released fix available in version 7.2.0 via GitHub PR #8610 and commit 68be1d12.

Syscall 12 (JumpToUser) in NovumOS versions prior to 0.24 executes arbitrary kernel code at Ring 0 privilege when invoked by unprivileged Ring 3 user-mode processes. The vulnerability stems from missing address validation on user-supplied entry points, enabling local privilege escalation from user mode to kernel mode with complete system control. CISA SSVC framework confirms publicly available exploit code (POC status) with total technical impact, though EPSS score remains low at 0.02% (5th percentile), suggesting limited real-world targeting of this niche custom operating system despite the severe technical flaw.

Symlink-based path traversal in the npm package 'compressing' v2.1.0 enables arbitrary file overwrites outside intended extraction directories via pre-planted symbolic links delivered through Git repositories. Attackers exploit a partial fix bypass of CVE-2026-24884 by poisoning filesystem state before archive extraction-Git clone operations automatically deploy malicious symlinks without user interaction beyond standard developer workflows. This supply chain vector allows overwriting critical system files (e.g., /etc/passwd) or application binaries to achieve privilege escalation or remote code execution. CVSS 8.4 (AV:L) reflects local attack vector, but real-world risk is amplified by Git-based delivery requiring zero privileges and no user interaction beyond cloning a malicious repository. No EPSS or KEV data available at time of analysis.

Command injection in Dolibarr ERP/CRM versions before 23.0.0 allows authenticated administrators to execute arbitrary operating system commands during ODT-to-PDF template conversion. The vulnerability stems from unsanitized concatenation of the MAIN_ODT_AS_PDF configuration constant into shell commands in odf.php. Exploitation requires administrative privileges (PR:H) but can be executed remotely (AV:N) with low complexity (AC:L), resulting in full system compromise as the web server user. Fixed in version 23.0.0. EPSS data not available; no public exploit identified at time of analysis.

Heap-based buffer overflow in xrdp 0.10.5 and earlier allows remote code execution after authentication via malicious EGFX graphics channel PDUs. Authenticated attackers can exploit insufficient validation of client-controlled size parameters to write beyond allocated heap buffers. Unauthenticated attackers can only trigger denial-of-service crashes. Vendor-released patch available in version 0.10.6. No active exploitation confirmed (not in CISA KEV), but heap overflows in remote services are high-value targets. Default non-privileged execution since 0.10.2 limits post-compromise impact.

Heap-based buffer overflow in xrdp's NeutrinoRDP module (versions ≤0.10.5) enables malicious downstream RDP servers or MitM attackers to achieve remote code execution or denial of service when proxying RDP sessions. Exploitation requires the victim xrdp server to have the non-default NeutrinoRDP module compiled and enabled (--enable-neutrinordp), and a user must initiate an RDP session through the affected proxy to a malicious server. EPSS data unavailable; no CISA KEV listing indicates targeted rather than widespread exploitation. Fixed in version 0.10.6.

Remote code execution in Anviz CX2 Lite and CX7 access control devices allows authenticated attackers to upload malicious firmware update packages that execute arbitrary scripts without verification. Reported by ICS-CERT, targeting physical access control systems commonly deployed in enterprise and critical infrastructure environments. CVSS 8.8 indicates high impact across confidentiality, integrity, and availability once low-privilege authentication is obtained. No public exploit confirmed at time of analysis, but the attack vector is straightforward for authenticated users.

Privilege escalation to root in xrdp 0.10.5 and earlier allows authenticated local attackers to execute arbitrary code due to improper error handling during privilege drop in the session execution component. The flaw requires low attack complexity and no user interaction (CVSS 8.8, AV:L/AC:L/PR:L/UI:N). Vendor-released patch available in xrdp v0.10.6. No public exploit or active exploitation confirmed at time of analysis, though CVSS scope change (S:C) indicates potential container/VM escape scenarios.

Remote code execution in Firebird RDBMS versions prior to 5.0.4, 4.0.7, and 3.0.14 allows authenticated users with CREATE FUNCTION privileges to execute arbitrary code as the database server process through path traversal in the external engine plugin loader. The vulnerability stems from insufficient input validation (CWE-22) when concatenating user-supplied engine names into filesystem paths, enabling attackers to load malicious shared libraries from arbitrary locations. With CVSS 10.0 and scope change (S:C), successful exploitation grants full system compromise beyond database boundaries. EPSS data not provided, no CISA KEV listing identified, indicating targeted rather than widespread exploitation at time of analysis. Vendor-released patches available across all affected major versions.

Remote code execution in Drag and Drop Multiple File Upload for Contact Form 7 plugin (WordPress) versions ≤1.3.9.6 allows unauthenticated attackers to upload PHP webshells via dual file validation weaknesses. The plugin's custom blacklist configuration overwrites default protections instead of merging, and non-ASCII filenames bypass the wpcf7_antiscript_file_name() sanitizer. CVSS 8.1 with High attack complexity (AV:N/AC:H/PR:N/UI:N). Wordfence reported; patch released in changeset 3508522. No KEV listing or confirmed public exploitation, but proof-of-concept feasible given detailed vulnerable code references (lines 62, 883, 970, 987).

Path traversal in WP Customer Area plugin through version 8.3.4 enables low-privileged authenticated users (Subscriber-level or higher, as configured by administrators) to read sensitive files like wp-config.php or delete critical WordPress files to achieve remote code execution. Wordfence reported this vulnerability affecting the ajax_attach_file function, which fails to properly validate file paths. CVSS 8.8 reflects network-based exploitation with low complexity, though exploitation requires authentication and administrator-granted plugin access. No evidence of active exploitation (not in CISA KEV) or public exploit code at time of analysis.

Quiz And Survey Master plugin for WordPress (versions up to 11.1.0) allows unauthenticated attackers to execute arbitrary WordPress shortcodes via user-submitted quiz answers. User inputs are sanitized with sanitize_text_field() and htmlspecialchars(), which strip HTML tags but fail to remove shortcode brackets [ and ]. When quiz results are displayed, the plugin executes do_shortcode() on the entire results page including user answers, enabling injection of shortcodes like [qsm_result id=X] to access unauthorized quiz submissions. This is a direct information disclosure vulnerability masked by RCE tagging; confirmed CVSS 5.3 (Integrity impact) indicates data tampering/unauthorized access rather than code execution.

Server-Side Request Forgery (SSRF) in Craftql PHP library versions 1.3.7 and earlier enables remote attackers to force the server to make unintended requests, potentially leading to arbitrary code execution. The vulnerability resides in the GetAssetsFieldSchema.php listener component. No active exploitation is confirmed (not in CISA KEV), but a proof-of-concept repository with detailed exploitation documentation exists on GitHub. Despite the CVSS 7.5 rating, the extremely low EPSS score (0.01%, 0th percentile) indicates minimal real-world exploitation activity observed to date. The description claims RCE capability, but the CVSS vector shows only confidentiality impact (C:H/I:N/A:N), suggesting the SSRF may enable information disclosure that could chain into RCE rather than direct code execution - verification with vendor advisories needed.

SiYuan is an open-source personal knowledge management system. In versions 3.6.3 and below, Mermaid diagrams are rendered with securityLevel set to "loose", and the resulting SVG is injected into the DOM via innerHTML. This allows attacker-controlled javascript: URLs in Mermaid code blocks to survive into the rendered output. On desktop builds using Electron, windows are created with nodeIntegration enabled and contextIsolation disabled, escalating the stored XSS to arbitrary code execution when a victim opens a note containing a malicious Mermaid block and clicks the rendered diagram node. This issue has been fixed in version 3.6.4.

SiYuan 3.6.1 through 3.6.3 allows arbitrary code execution when users view malicious bazaar packages in the marketplace UI. The vulnerability stems from an incomplete XSS fix (for CVE-2026-33066) that enabled an HTML sanitizer but failed to block iframe tags with srcdoc attributes containing embedded scripts. A malicious package author can inject JavaScript that executes in the Electron process with full application privileges, compromising the user's machine. The issue is confirmed fixed in version 3.6.4 and no public exploitation has been reported at time of analysis.

sagredo qmail before 2026.04.07 allows tls_quit remote code execution because of popen in notlshosts_auto in qmail-remote.c.

### Summary The webroot HTTP-01 challenge provider in lego is vulnerable to arbitrary file write and deletion via path traversal. A malicious ACME server can supply a crafted challenge token containing `../` sequences, causing lego to write attacker-influenced content to any path writable by the lego process. ### Details The `ChallengePath()` function in `challenge/http01/http_challenge.go:26-27` constructs the challenge file path by directly concatenating the ACME token without any validation: ```go func ChallengePath(token string) string { return "/.well-known/acme-challenge/" + token } ``` The webroot provider in `providers/http/webroot/webroot.go:31` then joins this with the configured webroot directory and writes the key authorization content to the resulting path: ```go challengeFilePath := filepath.Join(w.path, http01.ChallengePath(token)) err = os.MkdirAll(filepath.Dir(challengeFilePath), 0o755) err = os.WriteFile(challengeFilePath, []byte(keyAuth), 0o644) ``` RFC 8555 Section 8.3 specifies that ACME tokens must only contain characters from the base64url alphabet (`[A-Za-z0-9_-]`), but this constraint is never enforced anywhere in the codebase. When a malicious ACME server returns a token such as `../../../../../../tmp/evil`, `filepath.Join()` resolves the `..` components, producing a path outside the webroot directory. The same vulnerability exists in the `CleanUp()` function at `providers/http/webroot/webroot.go:48`, which deletes the challenge file using the same unsanitized path: ```go err := os.Remove(filepath.Join(w.path, http01.ChallengePath(token))) ``` This additionally enables arbitrary file deletion. ### PoC In a real attack scenario, the victim uses `--server` to point lego at a malicious ACME server, combined with `--http.webroot`: ```bash lego --server https://malicious-acme.example.com \ --http --http.webroot /var/www/html \ --email user@example.com \ --domains example.com \ run ``` The malicious server returns a challenge token containing path traversal sequences `../../../../../../tmp/pwned`. lego's webroot provider writes the key authorization to the traversed path without validation, resulting in arbitrary file write outside the webroot. The following minimal Go program demonstrates the core vulnerability by directly calling the webroot provider with a crafted token: ```go package main import ( "fmt" "os" "github.com/go-acme/lego/v4/providers/http/webroot" ) func main() { webrootDir, _ := os.MkdirTemp("", "lego-webroot-*") defer os.RemoveAll(webrootDir) provider, _ := webroot.NewHTTPProvider(webrootDir) token := "../../../../../../../../../../tmp/pwned" provider.Present("example.com", token, "EXPLOITED-BY-PATH-TRAVERSAL") data, err := os.ReadFile("/tmp/pwned") if err == nil { fmt.Println("[+] VULNERABILITY CONFIRMED") fmt.Printf("[+] File written outside webroot: /tmp/pwned\n") fmt.Printf("[+] Content: %s\n", data) } } ``` ```bash go build -o exploit ./exploit.go && ./exploit ``` Expected output: ``` [+] VULNERABILITY CONFIRMED [+] File written outside webroot: /tmp/pwned [+] Content: EXPLOITED-BY-PATH-TRAVERSAL ``` ### Impact This is a path traversal vulnerability (CWE-22). Any user running lego with the HTTP-01 challenge solver against a malicious or compromised ACME server is affected. A malicious ACME server can: - Achieve remote code execution by writing to cron directories, systemd unit paths, shell profiles, or web application directories served by the webroot. - Destroy data by overwriting configuration files, TLS certificates, or application state. - Escalate privileges if lego runs as root, granting unrestricted filesystem write access. - Delete arbitrary files via the `CleanUp()` code path using the same unsanitized token.

### Impact Up to 1.0.0 of `home-assitant-cli` (or `hass-cli` for short) an unrestricted environment was used to handle Jninja2 templates instead of a sandboxed one. The user-supplied input within Jinja2 templates was rendered locally with no restrictions. This gave users access to Python's internals and extended the scope of templating beyond the intended usage. E. g., it was possible to render a template with `hass-cli template bad-template.j2 --local` that contained entries like ````j2 {%- set b = environ.__globals__['__builtins__'] -%} {%- set os = b['__import__']('os') -%} {%- set bio = b['__import__']('builtins') -%} ... ```` or other malicious Jinja2 expressions. This can lead to arbitrary code execution on the local machine. In a two step process an adversary could trick/convince an user to download third-party templates which contain harmful code (e. g., perform data manipulation or establish a remote shell) then to render those templates unchecked/reviewed/verified with `--local`. The issue only affect the local machine and not a remote Home Assistant instance. It also requires user interventions. ### Patches 1.0.0 uses `ImmutableSandboxedEnvironment` and restricts the usage of environment variables. ### Workarounds Evaluate the Jninja2 templates manually or tool-based before rendering with `hass-cli`.

### Summary Due to unsafe serialization of stdio commands in the MCP adapter, an authenticated attacker can add an MCP stdio server with an arbitrary command, achieving command execution. ### Details The vulnerability lies in a bug in the input sanitization from the “Custom MCP” configuration in http://localhost:3000/canvas - where any user can add a new MCP, when doing so - adding a new MCP using stdio, the user can add any command, even though your code have input sanitization checks such as validateCommandInjection and validateArgsForLocalFileAccess, and a list of predefined specific safe commands - these commands, for example "npx" can be combined with code execution arguments ("-c touch /tmp/pwn") that enable direct code execution on the underlying OS. https://github.com/FlowiseAI/Flowise/blob/d848baeb6bd9737a1e7fc912349c45fbdcc7bb38/packages/components/nodes/tools/MCP/core.ts#L223 https://github.com/FlowiseAI/Flowise/blob/d848baeb6bd9737a1e7fc912349c45fbdcc7bb38/packages/components/nodes/tools/MCP/core.ts#L177 https://github.com/FlowiseAI/Flowise/blob/d848baeb6bd9737a1e7fc912349c45fbdcc7bb38/packages/components/nodes/tools/MCP/core.ts#L269 ### PoC Create a new Custom MCP and add an "npx -c" command. ``` { "command": "npx", "args": [ "-c", "touch /tmp/pwn" ] } ``` <img width="358" height="628" alt="Screenshot 2026-01-12 at 18 32 37" src="https://github.com/user-attachments/assets/d95c1ae2-23a7-4afe-b586-722003baf50e" /> ### Impact This is an authenticated arbitrary command execution due to unsanitized input, even though the input is sanitized, more protections should be added in order to close ways for attackers to execute arbitrary commands.

DataEase is an open-source data visualization and analytics platform. Versions 2.10.20 and below ship the legacy velocity-1.7.jar, which pulls in commons-collections-3.2.1.jar containing the InvokerTransformer deserialization gadget chain. Quartz 2.3.2, also bundled in the application, deserializes job data BLOBs from the qrtz_job_details table using ObjectInputStream with no deserialization filter or class allowlist. An authenticated attacker who can write to the Quartz job table, such as through the previously described SQL injection in previewSql, can replace a scheduled job's JOB_DATA with a malicious CommonsCollections6 gadget chain payload. When the Quartz cron trigger fires, the payload is deserialized and executes arbitrary commands as root inside the container, achieving full remote code execution. This issue has been fixed in version 2.10.21.

Incorrect use of boot service in the AMD Platform Configuration Blob (APCB) SMM driver could allow a privileged attacker with local access (Ring 0) to achieve privilege escalation potentially resulting in arbitrary code execution.

Improper validation of bash commands in Snowflake Cortex Code CLI versions prior to 1.0.25 allowed subsequent commands to execute outside the sandbox. An attacker could exploit this by embedding specially crafted commands in untrusted content, such as a malicious repository, causing the CLI agent to execute arbitrary code on the local device without user consent. Exploitation is non-deterministic and model-dependent. The fix is automatically applied upon relaunch with no user action required.

Remote code execution in Digital Knowledge KnowledgeDeliver (all versions prior to February 24, 2026) via malicious ViewState deserialization. A hard-coded ASP.NET machineKey allows unauthenticated remote attackers to bypass ViewState validation and execute arbitrary code on the server. Mandiant reported this critical deserialization vulnerability. EPSS score of 0.08% (24th percentile) suggests low observed exploitation activity, though no public exploit is confirmed at time of analysis. CVSS vector indicates network-accessible attack requiring no privileges or user interaction, but the 7.5 score reflects only Confidentiality impact-real-world RCE capability makes this significantly more severe than the partial CVSS rating suggests.