CyberHappenings logo
☰
Home Browse Watchlists Experimental About

Track cybersecurity events as they unfold. Sourced timelines. Filter, sort, and browse. Fast, privacy‑respecting. No invasive ads, no tracking.

Pre-authenticated RCE in Marimo exploited within 10 hours of advisory

First reported
Last updated
2 unique sources, 4 articles

Summary

Hide ▲

A pre-authenticated remote code execution (RCE) vulnerability in Marimo, tracked as CVE-2026-39987 (CVSS 9.3), was exploited in the wild within hours of public disclosure, enabling attackers to gain full PTY shells on exposed instances via the unauthenticated /terminal/ws WebSocket endpoint. The flaw, affecting Marimo versions prior to 0.23.0, initially led to rapid, human-driven exploitation campaigns focused on credential theft. Recent attacks have expanded to deploy a new NKAbuse malware variant hosted on Hugging Face Spaces, using typosquatted repositories to evade detection and establish persistent remote access. Additional exploitation activity includes sophisticated lateral movement, such as reverse-shell techniques and database enumeration, indicating a shift toward multi-stage attacks beyond initial credential harvesting. GitHub assessed the flaw with a critical CVSS score of 9.3, and Marimo developers confirmed impact on instances deployed as editable notebooks or exposed via --host 0.0.0.0 in edit mode. This event is distinct from the LMDeploy SSRF flaw (CVE-2026-33626), which was exploited within 12 hours of disclosure to conduct internal network reconnaissance via SSRF, targeting cloud metadata services, Redis, and MySQL instances.

Timeline

  1. 10.04.2026 10:37 4 articles · 14d ago

    CVE-2026-39987 exploited within 10 hours of Marimo advisory

    Within 9 hours and 41 minutes of public disclosure of CVE-2026-39987, threat actors exploited the pre-authenticated RCE in Marimo’s /terminal/ws WebSocket endpoint to obtain full PTY shells on exposed instances. Attackers manually explored compromised systems, retrieved .env files and SSH keys, and returned to confirm findings, indicating human-operated reconnaissance and data harvesting rather than automated payload deployment. New operational details show attackers validated the vulnerability within seconds of initial connection, completed credential harvesting in less than three minutes, and conducted reconnaissance from 125 IP addresses within the first 12 hours post-disclosure. The exploitation was part of a methodical, credential theft operation targeting .env files and SSH keys, with no attempt to install persistence or deploy automated payloads. Recent attacks now expand beyond credential theft, with attackers leveraging the Marimo RCE to deploy a new NKAbuse malware variant hosted on Hugging Face Spaces. The malware is delivered via a typosquatted Space (vsccode-modetx) hosting a dropper script (install-linux.sh) and a binary named kagent, which establishes persistence and provides remote access functionality over NKN decentralized network. Additional exploitation activity includes sophisticated lateral movement such as 15 reverse-shell techniques by a Germany-based actor and full Redis database dumps by a Hong Kong operator, indicating multi-stage attack campaigns beyond initial access. The expansion of tactics beyond credential theft and the deployment of persistent malware (e.g., NKAbuse variants) represent a significant evolution in exploitation patterns for this vulnerability.

    Show sources
    Open in new tab

Information Snippets

Similar Happenings

Active exploitation of maximum-severity code injection flaw in Flowise AI agent builder (CVE-2025-59528, CVSS 10.0)

Threat actors are actively exploiting CVE-2025-59528, a maximum-severity (CVSS 10.0) code injection vulnerability in Flowise, an open-source AI agent builder platform. The flaw allows unauthenticated remote code execution via the CustomMCP node, which parses user-supplied mcpServerConfig strings without sanitization, enabling execution of arbitrary JavaScript code with full Node.js runtime privileges. Successful exploitation grants attackers access to dangerous modules (e.g., child_process, fs), leading to full system compromise, arbitrary command execution, file system access, and sensitive data exfiltration. VulnCheck’s Canary network has detected early-stage exploitation of the flaw, originating from a single Starlink IP address. Between 12,000 and 15,000 Flowise instances remain exposed online, creating a significant and opportunistic attack surface for mass scanning and exploitation attempts. The vendor addressed the issue in versions 3.0.6 and later; users are advised to upgrade to at least 3.0.6 or the current 3.1.1.

Open in new tab

Langflow unauthenticated RCE vulnerability (CVE-2026-33017) exploited within 20 hours of disclosure

CISA formally confirmed active exploitation of the Langflow unauthenticated RCE vulnerability (CVE-2026-33017) on March 26, 2026, adding it to the Known Exploited Vulnerabilities (KEV) catalog and mandating U.S. federal agencies to apply mitigations or stop using the product by April 8, 2026. Threat actors exploited the flaw within 20–24 hours of its March 17, 2026 disclosure, progressing from automated scanning to staged Python payload delivery and credential harvesting (including .env and .db files) despite the absence of public PoC code. The vulnerability, with a CVSS score of 9.3, affects all Langflow versions prior to and including 1.8.1 and stems from an unsandboxed exec() call in the /api/v1/build_public_tmp/{flow_id}/flow endpoint. CISA did not attribute exploitation to ransomware actors but emphasized the risk to AI workflows given Langflow’s widespread adoption, including 145,000 GitHub stars. Endor Labs reported that attackers likely reverse-engineered exploits from the advisory details, underscoring the accelerating weaponization timeline. Mitigation guidance includes upgrading to version 1.9.0+ or disabling the vulnerable endpoint, restricting internet exposure, monitoring outbound traffic, and rotating all associated credentials.

Open in new tab

Critical Pre-Auth RCE Vulnerability in BeyondTrust Remote Support and PRA

BeyondTrust has patched a critical pre-authentication remote code execution (RCE) vulnerability (CVE-2026-1731, CVSS 9.9) in Remote Support (RS) and Privileged Remote Access (PRA) products. The flaw could allow unauthenticated attackers to execute OS commands in the context of the site user, leading to unauthorized access, data exfiltration, and service disruption. The vulnerability affects RS versions 25.3.1 and prior, and PRA versions 24.3.4 and prior. Patches are available in RS versions 25.3.2 and later, and PRA versions 25.1.1 and later. Self-hosted customers must manually apply updates if not subscribed to automatic updates. The vulnerability was discovered on January 31, 2026, with approximately 11,000 exposed instances identified, including around 8,500 on-prem deployments. BeyondTrust secured all RS/PRA cloud systems by February 2, 2026. The flaw was discovered by Harsh Jaiswal and the Hacktron AI team. Threat actors can exploit the flaw through maliciously crafted client requests in low-complexity attacks that do not require user interaction. In June 2025, BeyondTrust fixed a high-severity RS/PRA Server-Side Template Injection vulnerability. Attackers have begun actively exploiting the CVE-2026-1731 vulnerability in the wild, abusing the get_portal_info endpoint to extract the x-ns-company value before establishing a WebSocket channel. A proof-of-concept exploit targeting the /get_portal_info endpoint was published on GitHub. Threat actors have been observed exploiting CVE-2026-1731 to conduct network reconnaissance, deploy web shells, establish command-and-control (C2) channels, install backdoors and remote management tools, perform lateral movement, and exfiltrate data. The attacks have targeted financial services, legal services, high technology, higher education, wholesale and retail, and healthcare sectors across the U.S., France, Germany, Australia, and Canada. The vulnerability enables attackers to inject and execute arbitrary shell commands via the affected 'thin-scc-wrapper' script through the WebSocket interface. Attackers have deployed multiple web shells, including a PHP backdoor and a bash dropper, to maintain persistent access. Malware such as VShell and Spark RAT have been deployed as part of the exploitation. Out-of-band application security testing (OAST) techniques have been used to validate successful code execution and fingerprint compromised systems. Sensitive data, including configuration files, internal system databases, and a full PostgreSQL dump, have been exfiltrated to an external server. CVE-2026-1731 and CVE-2024-12356 share a common issue with input validation within distinct execution pathways. CVE-2026-1731 could be a target for sophisticated threat actors, similar to CVE-2024-12356 which was exploited by China-nexus threat actors like Silk Typhoon. CISA has confirmed that CVE-2026-1731 has been exploited in ransomware campaigns. CISA added CVE-2026-1731 to its Known Exploited Vulnerabilities (KEV) catalog on February 13 and gave federal agencies three days to apply the patch or stop using the product. Proof-of-concept (PoC) exploits for CVE-2026-1731 became available shortly after the initial disclosure, and exploitation was detected on January 31, making it a zero-day vulnerability for at least a week. CISA has activated the 'Known To Be Used in Ransomware Campaigns?' indicator in the KEV catalog for CVE-2026-1731. Customers of the cloud-based application (SaaS) had the patch applied automatically on February 2. Self-hosted instance customers need to either enable automatic updates or manually install the patch. For Remote Support, the recommended version is 25.3.2. For Privileged Remote Access, the recommended version is 25.1.1 or newer. Customers still using RS v21.3 and PRA v22.1 are recommended to upgrade to a newer version before applying the patch.

Open in new tab

Critical React Server Components (RSC) Bugs Enable Unauthenticated Remote Code Execution

A critical security vulnerability (CVE-2025-55182, CVSS 10.0) in React Server Components (RSC) allows unauthenticated remote code execution due to unsafe deserialization of payloads. The flaw affects multiple versions of React and Next.js, potentially impacting any application using RSC. The issue has been patched, but 39% of cloud environments remain vulnerable. Cloudflare experienced a widespread outage due to an emergency patch for this vulnerability, and multiple China-linked hacking groups have begun exploiting it. NHS England National CSOC has warned of the likelihood of continued exploitation in the wild. Major companies such as Google Cloud, AWS, and Cloudflare immediately responded to the vulnerability. The security researcher Lachlan Davidson disclosed the vulnerability on November 29, 2025, to the Meta team. The flaw has been dubbed React2Shell, a nod to the Log4Shell vulnerability discovered in 2021. The US National Vulnerability Database (NVD) rejected CVE-2025-66478 as a duplicate of CVE-2025-55182. Exploitation success rate is reported to be nearly 100% in default configurations. React servers that use React Server Function endpoints are known to be vulnerable. The Next.js web application is also vulnerable in its default configuration. At the time of writing, it is unknown if active exploitation has occurred, but there have been some reports of observed exploitation activity as of December 5, 2026. OX Security warned that the flaw is now actively exploitable on December 5, around 10am GMT. Hacker maple3142 published a working PoC, and OX Security successfully verified it. JFrog identified fake proof-of-concepts (PoC) on GitHub, warning security teams to verify sources before testing. Cloudflare started investigating issues on December 5 at 08:56 UTC, and a fix was rolled out within half an hour, but by that time outages had been reported by several major internet services, including Zoom, LinkedIn, Coinbase, DoorDash, and Canva. The U.S. Cybersecurity and Infrastructure Security Agency (CISA) added the vulnerability to its Known Exploited Vulnerabilities (KEV) catalog on December 6, 2025, following confirmed active exploitation. The vulnerability is tracked as React2Shell and is related to a remote code execution flaw in React Server Components (RSC). The flaw is due to insecure deserialization in the Flight protocol used by React to communicate between a server and client. The vulnerability affects versions 19.0, 19.1.0, 19.1.1, and 19.2.0 of react-server-dom-webpack, react-server-dom-parcel, and react-server-dom-turbopack. Patched versions of React are 19.0.1, 19.1.2, and 19.2.1. Downstream frameworks impacted include Next.js, React Router, Waku, Parcel, Vite, and RedwoodSDK. Amazon reported attack attempts from Chinese hacking groups like Earth Lamia and Jackpot Panda within hours of public disclosure. Coalition, Fastly, GreyNoise, VulnCheck, and Wiz reported seeing exploitation efforts targeting the flaw. Some attacks involved the deployment of cryptocurrency miners and the execution of "cheap math" PowerShell commands. Censys identified about 2.15 million instances of internet-facing services potentially affected by the vulnerability. Palo Alto Networks Unit 42 confirmed over 30 affected organizations across numerous sectors, with activity consistent with Chinese hacking group UNC5174. Security researcher Lachlan Davidson released multiple proof-of-concept (PoC) exploits for the vulnerability. Another working PoC was published by a Taiwanese researcher with the GitHub handle maple3142. Federal Civilian Executive Branch (FCEB) agencies have until December 26, 2025, to apply the necessary updates to secure their networks. Over 77,000 Internet-exposed IP addresses are vulnerable to the critical React2Shell remote code execution flaw (CVE-2025-55182). Researchers have confirmed that attackers have already compromised over 30 organizations across multiple sectors using the React2Shell flaw. Shadowserver detected 77,664 IP addresses vulnerable to the React2Shell flaw, with approximately 23,700 in the United States. GreyNoise recorded 181 distinct IP addresses attempting to exploit the flaw over the past 24 hours, with most of the traffic appearing automated. Attackers frequently begin with PowerShell commands that perform a basic math function to confirm the device is vulnerable to the remote code execution flaw. Once remote code execution was confirmed, attackers were seen executing base64-encoded PowerShell commands that download additional scripts directly into memory. One observed command executes a second-stage PowerShell script from the external site (23[.]235[.]188[.]3), which is used to disable AMSI to bypass endpoint security and deploy additional payloads. The PowerShell script observed by GreyNoise installs a Cobalt Strike beacon on the targeted device, giving threat actors a foothold on the network. Amazon AWS threat intelligence teams saw rapid exploitation hours after the disclosure of the React CVE-2025-55182 flaw, with infrastructure associated with China-linked APT hacking groups known as Earth Lamia and Jackpot Panda. Palo Alto Networks observed similar exploitation, attributing some of it to UNC5174, a Chinese state-sponsored threat actor believed to be tied to the Chinese Ministry of State Security. The deployed malware in these attacks includes Snowlight and Vshell, both commonly used by Chinese hacking groups for remote access, post-exploitation activity, and to move laterally through a compromised network. Earth Lamia is known for exploiting web application vulnerabilities to target organizations across Latin America, the Middle East, and Southeast Asia. Earth Lamia has historically targeted sectors across financial services, logistics, retail, IT companies, universities, and government organizations. Jackpot Panda primarily targets entities in East and Southeast Asia. The Shadowserver Foundation has identified over 77,000 vulnerable IPs following a scan of exposed HTTP services across a wide variety of exposed edge devices and other applications. Censys observed just over 2.15 million instances of internet-facing services that may be affected by this vulnerability, including exposed web services using React Server Components and exposed instances of frameworks such as Next.js, Waku, React Router, and RedwoodSDK. The bug is a pre-authentication remote code execution (RCE) vulnerability which exists in React Server Components versions 19.0.0, 19.1.0, 19.1.1, and 19.2.0. React issued a security advisory with the relevant patches and updates on December 3. Any internet-accessible server running the affected React Server Components code should be assumed vulnerable until updated as a precaution. AWS observed that many threat actors are attempting to use public PoCs that don’t work in real-world scenarios. AWS noted that the use of these PoCs shows that threat actors prioritize rapid operationalization over thorough testing, attempting to exploit targets with any available tool. Using multiple PoCs to scan for vulnerable environments also gives threat actors a higher chance of identifying vulnerable configurations, even if the PoCs are non-functional. The availability of the PoCs also allows less sophisticated actors to participate in exploitation campaigns. Finally, AWS noted that even failed exploitation attempts create significant noise in logs, potentially masking more sophisticated attacks. The invalid PoCs can give developers a false sense of security when testing for React2Shell. The Shadowserver Foundation detected 28,964 IP addresses vulnerable to the React2Shell flaw as of December 7, 2025, down from 77,664 on December 5, with approximately 10,100 located in the U.S., 3,200 in Germany, and 1,690 in China. Huntress observed attackers targeting numerous organizations via CVE-2025-55182, with a focus on the construction and entertainment industries. The first recorded exploitation attempt on a Windows endpoint by Huntress dates back to December 4, 2025, when an unknown threat actor exploited a vulnerable instance of Next.js to drop a shell script, followed by commands to drop a cryptocurrency miner and a Linux backdoor. Attackers were observed launching discovery commands and attempting to download several payloads from a command-and-control (C2) server. Huntress identified a Linux backdoor called PeerBlight, a reverse proxy tunnel named CowTunnel, and a Go-based post-exploitation implant referred to as ZinFoq. PeerBlight shares code overlaps with two malware families RotaJakiro and Pink that came to light in 2021, installs a systemd service to ensure persistence, and masquerades as a "ksoftirqd" daemon process to evade detection. CowTunnel initiates an outbound connection to attacker-controlled Fast Reverse Proxy (FRP) servers, effectively bypassing firewalls that are configured to only monitor inbound connections. ZinFoq implements a post-exploitation framework with interactive shell, file operations, network pivoting, and timestomping capabilities. Huntress assessed that the threat actor is likely leveraging automated exploitation tooling, supported by the attempts to deploy Linux-specific payloads on Windows endpoints, indicating the automation does not differentiate between target operating systems. PeerBlight supports capabilities to establish communications with a hard-coded C2 server ("185.247.224[.]41:8443"), allowing it to upload/download/delete files, spawn a reverse shell, modify file permissions, run arbitrary binaries, and update itself. ZinFoq beacons out to its C2 server and is equipped to parse incoming instructions to run commands using "/bin/bash," enumerate directories, read or delete files, download more payloads from a specified URL, exfiltrate files and system information, start/stop SOCKS5 proxy, enable/disable TCP port forwarding, alter file access and modification times, and establish a reverse pseudo terminal (PTY) shell connection. ZinFoq takes steps to clear bash history and disguises itself as one of 44 legitimate Linux system services to conceal its presence. CISA has urged federal agencies to patch the React2Shell vulnerability by December 12, 2025, amid reports of widespread exploitation. The vulnerability has been exploited by multiple threat actors in various campaigns to engage in reconnaissance efforts and deliver a wide range of malware families. Wiz observed a "rapid wave of opportunistic exploitation" of the flaw, with a vast majority of the attacks targeting internet-facing Next.js applications and other containerized workloads running in Kubernetes and managed cloud services. Cloudflare reported that threat actors have conducted searches using internet-wide scanning and asset discovery platforms to find exposed systems running React and Next.js applications. Some of the reconnaissance efforts have excluded Chinese IP address spaces from their searches. The observed activity targeted government (.gov) websites, academic research institutions, and critical-infrastructure operators. Early scanning and exploitation attempts originated from IP addresses previously associated with Asia-affiliated threat clusters. Kaspersky recorded over 35,000 exploitation attempts on a single day on December 10, 2025, with the attackers first probing the system by running commands like whoami, before dropping cryptocurrency miners or botnet malware families like Mirai/Gafgyt variants and RondoDox. Security researcher Rakesh Krishnan discovered an open directory hosted on "154.61.77[.]105:8082" that includes a proof-of-concept (PoC) exploit script for CVE-2025–55182 along with two other files: "domains.txt," which contains a list of 35,423 domains, and "next_target.txt," which contains a list of 596 URLs, including companies like Dia Browser, Starbucks, Porsche, and Lululemon. The Shadowserver Foundation reported more than 137,200 internet-exposed IP addresses running vulnerable code as of December 11, 2025, with over 88,900 instances located in the U.S., followed by Germany (10,900), France (5,500), and India (3,600). Google's threat intelligence team linked five more Chinese hacking groups to attacks exploiting the React2Shell vulnerability. The list of state-linked threat groups exploiting the flaw now also includes UNC6600, UNC6586, UNC6588, UNC6603, and UNC6595. GTIG researchers observed numerous discussions regarding CVE-2025-55182 in underground forums, including threads where threat actors shared links to scanning tools, proof-of-concept (PoC) code, and their experiences using these tools. GTIG also spotted Iranian threat actors targeting the flaw and financially motivated attackers deploying XMRig cryptocurrency mining software on unpatched systems. Shadowserver Internet watchdog group is currently tracking over 116,000 IP addresses vulnerable to React2Shell attacks, with over 80,000 in the United States. GreyNoise has observed over 670 IP addresses attempting to exploit the React2Shell remote code execution flaw over the past 24 hours, primarily originating from the United States, India, France, Germany, the Netherlands, Singapore, Russia, Australia, the United Kingdom, and China. Threat actors are exploiting the React2Shell vulnerability to deliver malware families like KSwapDoor and ZnDoor. KSwapDoor is a professionally engineered remote access tool designed with stealth in mind, building an internal mesh network and using military-grade encryption. KSwapDoor impersonates a legitimate Linux kernel swap daemon to evade detection. ZnDoor is a remote access trojan that contacts threat actor-controlled infrastructure to receive and execute commands. ZnDoor supports commands such as shell, interactive_shell, explorer, explorer_cat, explorer_delete, explorer_upload, explorer_download, system, change_timefile, socket_quick_startstreams, start_in_port_forward, and stop_in_port. Google identified five China-nexus groups exploiting React2Shell to deliver various payloads, including MINOCAT, SNOWLIGHT, COMPOOD, HISONIC, and ANGRYREBEL. Microsoft reported that threat actors have used the flaw to run arbitrary commands, set up reverse shells, drop RMM tools, and modify authorized_keys files. Payloads delivered in these attacks include VShell, EtherRAT, SNOWLIGHT, ShadowPad, and XMRig. Threat actors used Cloudflare Tunnel endpoints to evade security defenses and conducted reconnaissance for lateral movement and credential theft. Credential harvesting targeted Azure Instance Metadata Service (IMDS) endpoints for Azure, AWS, GCP, and Tencent Cloud. Threat actors deployed secret discovery tools such as TruffleHog and Gitleaks, along with custom scripts to extract various secrets. Beelzebub detailed a campaign exploiting Next.js flaws to extract credentials and sensitive data, including environment files, SSH keys, cloud credentials, and system files. The malware creates persistence, installs a SOCKS5 proxy, establishes a reverse shell, and installs a React scanner for further propagation. Operation PCPcat has breached an estimated 59,128 servers. The Shadowserver Foundation is tracking over 111,000 IP addresses vulnerable to React2Shell attacks, with over 77,800 instances in the U.S. GreyNoise observed 547 malicious IP addresses from the U.S., India, the U.K., Singapore, and the Netherlands partaking in exploitation efforts over the past 24 hours. The RondoDox botnet has been observed exploiting the critical React2Shell flaw (CVE-2025-55182) to infect vulnerable Next.js servers with malware and cryptominers. First documented by Fortinet in July 2025, RondoDox is a large-scale botnet that targets multiple n-day flaws in global attacks. In November, VulnCheck spotted new RondoDox variants that featured exploits for CVE-2025-24893, a critical remote code execution (RCE) vulnerability in the XWiki Platform. A new report from cybersecurity company CloudSEK notes that RondoDox started scanning for vulnerable Next.js servers on December 8 and began deploying botnet clients three days later. React2Shell is an unauthenticated remote code execution vulnerability that can be exploited via a single HTTP request and affects all frameworks that implement the React Server Components (RSC) 'Flight' protocol, including Next.js. The flaw has been leveraged by several threat actors to breach multiple organizations. North Korean hackers exploited React2Shell to deploy a new malware family named EtherRAT. As of December 30, the Shadowserver Foundation reports detecting over 94,000 internet-exposed assets vulnerable to React2Shell. CloudSEK says that RondoDox has passed through three distinct operational phases this year: Reconnaissance and vulnerability testing from March to April 2025, Automated web app exploitation from April to June 2025, Large-scale IoT botnet deployment from July to today. Regarding React2Shell, the researchers report that RondoDox has focused its exploitation around the flaw significantly lately, launching over 40 exploit attempts within six days in December. During this operational phase, the botnet conducts hourly IoT exploitation waves targeting Linksys, Wavlink, and other consumer and enterprise routers to enroll new bots. After probing potentially vulnerable servers, CloudSEK says that RoundDox started to deploy payloads that included a coinminer (/nuts/poop), a botnet loader and health checker (/nuts/bolts), and a variant of Mirai (/nuts/x86). The 'bolts' component removes competing botnet malware from the host, enforces persistence via /etc/crontab, and kills non-whitelisted processes every 45 seconds, the researchers say. CloudSEK provides a set of recommendations for companies to protect against this RondoDox activity, among them auditing and patching Next.js Server Actions, isolating IoT devices into dedicated virtual LANs, and monitoring for suspicious processes being executed. Threat actors targeting cloud environments now favor campaigns which gain initial access by exploiting software vulnerabilities over credential-based attacks. Third-party software-based entry accounted for 44.5% of primary entry vectors during the second half of 2025, up from 2.9% in the first half. Abuse of weak or absent credentials as an entry point dropped from 47.1% in the first half of 2025 to 27.2% in the second half. React2Shell (CVE-2025-55182) was one of the most commonly exploited vulnerabilities to target cloud services. Google Cloud noted that within 48 hours of the public disclosure of React2Shell, multiple threat actors had already exploited the vulnerability to infect victims with cryptocurrency mining malware. The window between vulnerability disclosure and mass exploitation collapsed from weeks to just days. Google Cloud recommended using centralized visibility tools to secure data and automated posture enforcement to mitigate risks. Google Cloud advised organizations to pivot from manual patching to automated defenses, such as patching the Web Application Firewall (WAF), to neutralize exploits at the network edge before software updates can be applied.

Open in new tab

React Native CLI Remote Code Execution Vulnerability (CVE-2025-11953)

A critical security flaw in the React Native CLI package, tracked as CVE-2025-11953, allowed remote, unauthenticated attackers to execute arbitrary OS commands on development servers. The vulnerability affected versions 4.8.0 through 20.0.0-alpha.2 of the @react-native-community/cli-server-api package, impacting millions of developers using the React Native framework. The flaw was patched in version 20.0.0. The vulnerability is being actively exploited in the wild, with attacks observed on December 21, 2025, January 4, 2026, and January 21, 2026. The attacks involve delivering base-64 encoded PowerShell payloads hidden in the HTTP POST body of malicious requests. The payloads disable endpoint protections, establish a raw TCP connection to attacker-controlled infrastructure, write data to disk, and execute the downloaded binary. Approximately 3,500 exposed React Native Metro servers are still online, according to scans using the ZoomEye search engine. Despite active exploitation being observed for over a month, the vulnerability still carries a low score in the Exploit Prediction Scoring System (EPSS). The vulnerability affects Windows, Linux, and macOS systems, with varying levels of control over executed commands. The flaw was discovered by researchers at JFrog and disclosed in early November 2025. The vulnerability is dubbed Metro4Shell by VulnCheck. The Windows payload is a Rust-based UPX-packed binary with basic anti-analysis logic, and the same attacker infrastructure hosts corresponding Linux binaries, indicating cross-platform targeting.

Open in new tab