Roo Code is an AI-powered autonomous coding agent. The project-specific MCP configuration for the Roo Code agent is stored in the `.roo/mcp.json` file within the VS Code workspace. Because the MCP configuration format allows for execution of arbitrary commands, prior to version 3.20.3, it would have been possible for an attacker with access to craft a prompt to ask the agent to write a malicious command to the MCP configuration file. If the user had opted-in to auto-approving file writes within the project, this would have led to arbitrary command execution. This issue is of moderate severity, since it requires the attacker to already be able to submit prompts to the agent (for instance through a prompt injection attack), for the user to have MCP enabled (on by default), and for the user to have enabled auto-approved file writes (off by default). Version 3.20.3 fixes the issue by adding an additional layer of opt-in configuration for auto-approving writing to Roo's configuration files, including all files within the `.roo/` folder.

Roo Code is an AI-powered autonomous coding agent that lives in users' editors. Versions 3.25.23 and below contain a vulnerability where certain VS Code workspace configuration files (.code-workspace) are not protected in the same way as the .vscode folder. If the agent was configured to auto-approve file writes, an attacker able to influence prompts (for example via prompt injection) could cause malicious workspace settings or tasks to be written. These tasks could then be executed automatically when the workspace is reopened, resulting in arbitrary code execution. This issue is fixed in version 3.26.0.

Roo Code is an AI-powered autonomous coding agent that lives in users' editors. Versions below 3.26.0 contain a vulnerability in the command parsing logic where the Bash parameter expansion and indirect reference were not handled correctly. If the agent was configured to auto-approve execution of certain commands, an attacker able to influence prompts could abuse this weakness to execute additional arbitrary commands alongside the intended one. This is fixed in version 3.26.0.

Roo Code is an AI-powered autonomous coding agent that lives in users' editors. In versions prior to 3.25.5, Roo-Code fails to properly handle process substitution and single ampersand characters in the command parsing logic for auto-execute commands. If a user has enabled auto-approved execution for a command such as ls, an attacker who can submit crafted prompts to the agent may inject arbitrary commands to be executed alongside the intended command. Exploitation requires attacker access to submit prompts and for the user to have enabled auto-approved command execution, which is disabled by default. This vulnerability could allow an attacker to execute arbitrary code. The issue is fixed in version 3.25.5.

Roo Code is an AI-powered autonomous coding agent. Prior to 3.22.6, if the victim had "Write" auto-approved, an attacker with the ability to submit prompts to the agent could write to VS Code settings files and trigger code execution. There were multiple ways to achieve that. One example is with the php.validate.executablePath setting which lets you set the path for the php executable for syntax validation. The attacker could have written the path to an arbitrary command there and then created a php file to trigger it. This vulnerability is fixed in 3.22.6.

Roo Code is an AI-powered autonomous coding agent that lives in users' editors. In versions 3.23.18 and below, RooCode does not validate line breaks (\n) in its command input, allowing potential bypass of the allow-list mechanism. The project appears to lack parsing or validation logic to prevent multi-line command injection. When commands are evaluated for execution, only the first line or token may be considered, enabling attackers to smuggle additional commands in subsequent lines. This is fixed in version 3.23.19.

The LearnPress plugin for WordPress is vulnerable to Command Injection in all versions up to, and including, 4.2.5.7 via the get_content function. This is due to the plugin making use of the call_user_func function with user input. This makes it possible for unauthenticated attackers to execute any public function with one parameter, which could result in remote code execution.

A vulnerability, which was classified as critical, was found in Zarthus IRC Twitter Announcer Bot up to 1.1.0. This affects the function get_tweets of the file lib/twitterbot/plugins/twitter_announcer.rb. The manipulation of the argument tweet leads to command injection. It is possible to initiate the attack remotely. The complexity of an attack is rather high. The exploitability is told to be difficult. Upgrading to version 1.1.1 is able to address this issue. The patch is named 6b1941b7fc2c70e1f40981b43c84a2c20cc12bd3. It is recommended to upgrade the affected component. The associated identifier of this vulnerability is VDB-223383.

Woodpecker is a community fork of the Drone CI system. In affected versions an attacker can post malformed webhook data witch lead to an update of the repository data that can e.g. allow the takeover of an repo. This is only critical if the CI is configured for public usage and connected to a forge witch is also in public usage. This issue has been addressed in version 1.0.2. Users are advised to upgrade. Users unable to upgrade should secure the CI system by making it inaccessible to untrusted entities, for example, by placing it behind a firewall.

VMware Aria Operations contains a command injection vulnerability. A malicious unauthenticated actor may exploit this issue to execute arbitrary commands which may lead to remote code execution in VMware Aria Operations while support-assisted product migration is in progress.  To remediate CVE-2026-22719, apply the patches listed in the 'Fixed Version' column of the ' Response Matrix https://support.broadcom.com/web/ecx/support-content-notification/-/external/content/SecurityAdvisories/0/36947 ' in VMSA-2026-0001  Workarounds for CVE-2026-22719 are documented in the 'Workarounds' column of the ' Response Matrix https://support.broadcom.com/web/ecx/support-content-notification/-/external/content/SecurityAdvisories/0/36947 ' in VMSA-2026-0001

Deno is a JavaScript, TypeScript, and WebAssembly runtime. Before 2.5.6, a prior patch aimed to block spawning Windows batch/shell files by returning an error when a spawned path’s extension matched .bat or .cmd. That check performs a case-sensitive comparison against lowercase literals and therefore can be bypassed when the extension uses alternate casing (for example .BAT, .Bat, etc.). This vulnerability is fixed in 2.5.6.

All versions of package gitpython are vulnerable to Remote Code Execution (RCE) due to improper user input validation, which makes it possible to inject a maliciously crafted remote URL into the clone command. Exploiting this vulnerability is possible because the library makes external calls to git without sufficient sanitization of input arguments.

Frourio is a full stack framework, for TypeScript. Frourio users who uses frourio version prior to v0.26.0 and integration with class-validator through `validators/` folder are subject to a input validation vulnerability. Validators do not work properly for request bodies and queries in specific situations and some input is not validated at all. Users are advised to update frourio to v0.26.0 or later and to install `class-transformer` and `reflect-metadata`.

A vulnerability was found in ChurchCRM up to 5.18.0. This vulnerability affects unknown code of the file setup/routes/setup.php. Performing a manipulation of the argument DB_PASSWORD/ROOT_PATH/URL results in deserialization. The attack may be initiated remotely. The attack's complexity is rated as high. It is stated that the exploitability is difficult. The exploit has been made public and could be used. The vendor was contacted early about this disclosure but did not respond in any way.

Frourio-express is a minimal full stack framework, for TypeScript. Frourio-express users who uses frourio-express version prior to v0.26.0 and integration with class-validator through `validators/` folder are subject to a input validation vulnerability. Validators do not work properly for request bodies and queries in specific situations and some input is not validated at all. Users are advised to update frourio to v0.26.0 or later and to install `class-transformer` and `reflect-metadata`.

Server receiving a malformed message to create a new connection could lead to an attacker performing remote code execution or causing a failure. See Honeywell Security Notification for recommendations on upgrading and versioning.

An Improper Input Validation vulnerability in the J-Web component of Juniper Networks Junos OS may allow an unauthenticated attacker to access data without proper authorization. Utilizing a crafted POST request, deserialization may occur which could lead to unauthorized local file access or the ability to execute arbitrary commands. This issue affects Juniper Networks Junos OS: all versions prior to 19.1R3-S9; 19.2 versions prior to 19.2R3-S6; 19.3 versions prior to 19.3R3-S7; 19.4 versions prior to 19.4R2-S7, 19.4R3-S9; 20.1 versions prior to 20.1R3-S5; 20.2 versions prior to 20.2R3-S5; 20.3 versions prior to 20.3R3-S5; 20.4 versions prior to 20.4R3-S4; 21.1 versions prior to 21.1R3-S2; 21.2 versions prior to 21.2R3-S1; 21.3 versions prior to 21.3R2-S2, 21.3R3; 21.4 versions prior to 21.4R1-S2, 21.4R2-S1, 21.4R3; 22.1 versions prior to 22.1R1-S1, 22.1R2.

A vulnerability was found in MicroWorld eScan Antivirus 7.0.32 on Linux. It has been rated as critical. This issue affects some unknown processing of the file rtscanner of the component Quarantine Handler. The manipulation leads to os command injection. The attack may be initiated remotely. The complexity of an attack is rather high. The exploitation is known to be difficult. The exploit has been disclosed to the public and may be used. The vendor was contacted early about this disclosure but did not respond in any way.

Improper Neutralization of Special Elements used in a Command ('Command Injection') vulnerability in Xiaomi Xiaomi Router allows Command Injection.

Certain Lexmark devices through 2023-02-19 mishandle Input Validation (issue 1 of 4).

This affects versions of the package pydash before 6.0.0. A number of pydash methods such as pydash.objects.invoke() and pydash.collections.invoke_map() accept dotted paths (Deep Path Strings) to target a nested Python object, relative to the original source object. These paths can be used to target internal class attributes and dict items, to retrieve, modify or invoke nested Python objects. **Note:** The pydash.objects.invoke() method is vulnerable to Command Injection when the following prerequisites are satisfied: 1) The source object (argument 1) is not a built-in object such as list/dict (otherwise, the __init__.__globals__ path is not accessible) 2) The attacker has control over argument 2 (the path string) and argument 3 (the argument to pass to the invoked method) The pydash.collections.invoke_map() method is also vulnerable, but is harder to exploit as the attacker does not have direct control over the argument to be passed to the invoked function.

A vulnerability was found in Monitorr up to 1.7.6m. It has been classified as problematic. This affects an unknown part of the file assets/config/_installation/mkdbajax.php of the component Installer. The manipulation of the argument datadir leads to improper input validation. It is possible to initiate the attack remotely. The complexity of an attack is rather high. The exploitability is told to be difficult. The exploit has been disclosed to the public and may be used. The vendor was contacted early about this disclosure but did not respond in any way.

A vulnerability, which was classified as critical, has been found in Comodo Internet Security Premium 12.3.4.8162. This issue affects some unknown processing of the file cis_update_x64.xml of the component Manifest File Handler. The manipulation of the argument binary/params leads to os command injection. The attack may be initiated remotely. The complexity of an attack is rather high. The exploitation is known to be difficult. The exploit has been disclosed to the public and may be used. The vendor was contacted early about this disclosure but did not respond in any way.

An os command injection vulnerability exists in the liburvpn.so create_private_key functionality of Milesight VPN v2.0.2. A specially-crafted network request can lead to command execution. An attacker can send a malicious packet to trigger this vulnerability.

It was found that the original fix for log4j CVE-2021-44228 and CVE-2021-45046 in the OpenShift metering hive containers was incomplete, as not all JndiLookup.class files were removed. This CVE only applies to the OpenShift Metering hive container images, shipped in OpenShift 4.8, 4.7 and 4.6.

Deno is a JavaScript, TypeScript, and WebAssembly runtime. Versions prior to 2.5.3 and 2.2.15 are vulnerable to Command Line Injection attacks on Windows when batch files are executed. In Windows, ``CreateProcess()`` always implicitly spawns ``cmd.exe`` if a batch file (.bat, .cmd, etc.) is being executed even if the application does not specify it via the command line. This makes Deno vulnerable to a command injection attack on Windows. Versions 2.5.3 and 2.2.15 fix the issue.

Netdata is an open source option for real-time infrastructure monitoring and troubleshooting. An attacker with the ability to establish a streaming connection can execute arbitrary commands on the targeted Netdata agent. When an alert is triggered, the function `health_alarm_execute` is called. This function performs different checks and then enqueues a command by calling `spawn_enq_cmd`. This command is populated with several arguments that are not sanitized. One of them is the `registry_hostname` of the node for which the alert is raised. By providing a specially crafted `registry_hostname` as part of the health data that is streamed to a Netdata (parent) agent, an attacker can execute arbitrary commands at the remote host as a side-effect of the raised alert. Note that the commands are executed as the user running the Netdata Agent. This user is usually named `netdata`. The ability to run arbitrary commands may allow an attacker to escalate privileges by escalating other vulnerabilities in the system, as that user. The problem has been fixed in: Netdata agent v1.37 (stable) and Netdata agent v1.36.0-409 (nightly). As a workaround, streaming is not enabled by default. If you have previously enabled this, it can be disabled. Limiting access to the port on the recipient Agent to trusted child connections may mitigate the impact of this vulnerability.

Improper input validation in Windows Server Update Service allows an unauthorized attacker to execute code over a network.

cn.wps.moffice.common.beans.print.CloudPrintWebView in Kingsoft Office 5.3.1, as used in Huawei P2 devices before V100R001C00B043, falls back to HTTP when the HTTPS connection to the registry fails, which allows man-in-the-middle attackers to conduct downgrade attacks and execute arbitrary Java code by leveraging a network position between the client and the registry to block HTTPS traffic.

A vulnerability exists in EnOcean SmartServer IoT version 4.60.009 and prior, which would allow remote attackers, in the LON IP-852 management messages, to send specially crafted IP-852 messages resulting in arbitrary OS command execution on the device.

A vulnerability was found in Jrohy trojan up to 2.15.3. It has been declared as critical. This vulnerability affects the function LogChan of the file trojan/util/linux.go. The manipulation of the argument c leads to os command injection. The attack can be initiated remotely. The complexity of an attack is rather high. The exploitation appears to be difficult. The exploit has been disclosed to the public and may be used.

A flaw has been found in wong2 mcp-cli 1.13.0. Affected is the function redirectToAuthorization of the file /src/oauth/provider.js of the component oAuth Handler. This manipulation causes os command injection. The attack may be initiated remotely. The attack is considered to have high complexity. The exploitability is told to be difficult. The exploit has been published and may be used. The vendor was contacted early about this disclosure but did not respond in any way.

Improper Limitation of a Pathname to a Restricted Directory ('Path Traversal'), Improper Neutralization of Special Elements used in a Command ('Command Injection') vulnerability in flow computer and remote controller products of ABB ( RMC-100 (Standard), RMC-100-LITE, XIO, XFCG5 , XRCG5 , uFLOG5 , UDC) allows an attacker who successfully exploited this vulnerability could insert and run arbitrary code in an affected system node.

PrinterLogic Web Stack versions 19.1.1.13 SP9 and below do not sanitize user input resulting in pre-auth remote code execution.

totolink EX300_v2 V4.0.3c.140_B20210429 was discovered to contain a command injection vulnerability via the component process forceugpo.

AIMANAGER before B115 on MONITORAPP Application Insight Web Application Firewall (AIWAF) devices with Manager 2.1.0 allows OS Command Injection because of missing input validation on one of the parameters of an HTTP request.

An issue was discovered in dhclient 4.3.1-6 due to an embedded path variable.

This issues due to insufficient verification of the various input values from user’s input. The vulnerability allows remote attackers to execute malicious code in Firstmall via navercheckout_add function.

In Go before 1.10.6 and 1.11.x before 1.11.3, the "go get" command is vulnerable to directory traversal when executed with the import path of a malicious Go package which contains curly braces (both '{' and '}' characters). Specifically, it is only vulnerable in GOPATH mode, but not in module mode (the distinction is documented at https://golang.org/cmd/go/#hdr-Module_aware_go_get). The attacker can cause an arbitrary filesystem write, which can lead to code execution.

In Go before 1.10.6 and 1.11.x before 1.11.3, the "go get" command is vulnerable to remote code execution when executed with the -u flag and the import path of a malicious Go package, or a package that imports it directly or indirectly. Specifically, it is only vulnerable in GOPATH mode, but not in module mode (the distinction is documented at https://golang.org/cmd/go/#hdr-Module_aware_go_get). Using custom domains, it's possible to arrange things so that a Git repository is cloned to a folder named ".git" by using a vanity import path that ends with "/.git". If the Git repository root contains a "HEAD" file, a "config" file, an "objects" directory, a "refs" directory, with some work to ensure the proper ordering of operations, "go get -u" can be tricked into considering the parent directory as a repository root, and running Git commands on it. That will use the "config" file in the original Git repository root for its configuration, and if that config file contains malicious commands, they will execute on the system running "go get -u".

Ansible before versions 2.1.4, 2.2.1 is vulnerable to an improper input validation in Ansible's handling of data sent from client systems. An attacker with control over a client system being managed by Ansible and the ability to send facts back to the Ansible server could use this flaw to execute arbitrary code on the Ansible server using the Ansible server privileges.

The SMBv1 server in Microsoft Windows Vista SP2; Windows Server 2008 SP2 and R2 SP1; Windows 7 SP1; Windows 8.1; Windows Server 2012 Gold and R2; Windows RT 8.1; and Windows 10 Gold, 1511, and 1607; and Windows Server 2016 allows remote attackers to execute arbitrary code via crafted packets, aka "Windows SMB Remote Code Execution Vulnerability." This vulnerability is different from those described in CVE-2017-0143, CVE-2017-0144, CVE-2017-0145, and CVE-2017-0146.

A flaw was found in the way samba before 4.7.9 and 4.8.4 allowed the use of weak NTLMv1 authentication even when NTLMv1 was explicitly disabled. A man-in-the-middle attacker could use this flaw to read the credential and other details passed between the samba server and client.

1Panel is a web interface and MCP Server that manages websites, files, containers, databases, and LLMs on a Linux server. In versions 2.0.5 and below, the HTTPS protocol used for communication between the Core and Agent endpoints has incomplete certificate verification during certificate validation, leading to unauthorized interface access. Due to the presence of numerous command execution or high-privilege interfaces in 1Panel, this results in Remote Code Execution (RCE). This is fixed in version 2.0.6. The CVE has been translated from Simplified Chinese using GitHub Copilot.

Successful exploitation of the vulnerability could allow an attacker to execute arbitrary commands as root, potentially leading to the loss of confidentiality, integrity, availability, and full control of the access point.

A vulnerability was found in Ackites KillWxapkg up to 2.4.1. It has been declared as critical. This vulnerability affects the function processFile of the file internal/unpack/unpack.go of the component wxapkg File Parser. The manipulation leads to os command injection. The attack can be initiated remotely. The complexity of an attack is rather high. The exploitation appears to be difficult. The exploit has been disclosed to the public and may be used.

An issue was discovered in the slock crate through 2020-11-17 for Rust. Slock<T> unconditionally implements Send and Sync.

An issue was discovered in the dces crate through 2020-12-09 for Rust. The World type is marked as Send but lacks bounds on its EntityStore and ComponentStore.

A race condition exists in PaperCut MF when processing badge-swipe data from certain HP multifunction devices. Under specific network conditions involving dropped packets and out-of-order sequence counters, the server may incorrectly process fragmented data chunks. If a sequence reset notification fails to reach the server, the server may reject the initial data chunk while erroneously accepting subsequent chunks before a connection reset completes. This leads to the registration of a truncated badge ID string. While this typically results in an authentication failure, the vulnerability is compounded in environments utilizing custom badge-ID post-processing scripts. In such configurations, the truncated string may be transformed into a valid ID belonging to a different user, leading to unauthorized session establishment (Incorrect User Login) on the device.