Attackers exploit critical Flowise flaw CVE-2025
Vulnerabilities mentionedAll →
| CVE | Vulnerability | CVSS | EPSS | Flags | Affected | Exposure | Published |
|---|---|---|---|---|---|---|---|
| CVE-2025-26319 | Unauthenticated Arbitrary File Upload in FlowiseAI Flowise 2.2.6 CVE-2025-26319 is an arbitrary file upload flaw (CWE-434) in the /api/v1/attachments API endpoint of FlowiseAI Flowise v2.2.6, a self-hosted open-source platform for building AI agent and LLM workflows. The endpoint accepts uploads without proper validation, and because the flaw is network-reachable with no authentication and no user interaction required (per the CVSS vector), a remote unauthenticated attacker can trigger it by sending crafted upload requests. The critical 9.8 score with high ratings for confidentiality, integrity, and availability indicates that successful abuse can lead to full compromise of the host, plausibly including remote code execution via uploaded files. Anyone running Flowise v2.2.6 is affected, and the risk is highest for instances exposed to the internet. Exploitation appears to be ongoing: a public proof-of-concept is available on GitHub, EPSS puts the 30-day exploitation probability at 55.9% (99th percentile), and security news reports attackers actively exploiting a critical Flowise flaw against the product's roughly 12,000+ internet-exposed instances. Do: Upgrade Flowise to the latest release beyond v2.2.6 (no specific fixed version is stated in the available data, so verify the fix with the vendor's changelog), and keep internet exposure of /api/v1/attachments restricted — place Flowise behind a reverse proxy with authentication or limit it to trusted networks. Review servers for unexpected or recently created uploaded files and check access logs for unauthenticated POSTs to /api/v1/attachments as evidence of exploitation. | 9.8 | 56% | PoC |
| large≈12,000+ internet-exposed Flowise instances (public scan reporting) | |
| CVE-2025-59528 | Critical Unauthenticated RCE in Flowise AI Agent Builder Flowise, an open-source drag-and-drop interface for building LLM agent flows, contains a critical remote code execution flaw (CWE-94, code injection) in its CustomMCP node. The node parses the user-supplied mcpServerConfig string, and the convertToValidJSONString helper passes this input directly to the JavaScript Function() constructor without any validation, causing it to be evaluated and executed as live code. Because the evaluation runs with full Node.js runtime privileges, an attacker can invoke dangerous modules such as child_process and fs to execute arbitrary commands on the host. Any Flowise deployment running version 3.0.5 or earlier is affected, and the CVSS vector (network attack vector, no privileges or user interaction required, scope change) indicates unauthenticated remote attackers can trigger it. Exploitation is active in the wild according to news reports citing roughly 12,000 exposed instances, and EPSS assigns an 87.4% probability of exploitation within 30 days. Do: Upgrade Flowise to version 3.0.6 or later immediately. Given confirmed in-the-wild exploitation, review internet-exposed instances for signs of compromise (unexpected processes, spawned shells, or modified files) and restrict access to CustomMCP-configured instances behind authentication or a VPN until patched. The flaw is not yet listed in CISA KEV, but defenders should treat patching as urgent given the CVSS 10.0 rating and 87.4% EPSS score. | 10.0 | 86% | PoC |
| large~12,000+ internet-exposed Flowise instances | |
| CVE-2025-8943 | Unauthenticated Remote Code Execution in Flowise AI Agent Builder Flowise's Custom MCPs feature launches local MCP servers by executing operating-system commands (for example via npx), but Flowise has a minimal authentication and authorization model with no role-based access controls, and versions before 3.0.1 run without authentication by default unless it is explicitly configured. As a result, an unauthenticated attacker with network access to an affected instance can invoke this functionality to run unsandboxed OS commands. Successful exploitation yields full unauthenticated remote code execution on the host with the privileges of the Flowise service, with high impact to confidentiality, integrity, and availability (CVSS 3.1: 9.8). Any Flowise deployment prior to 3.0.1 is affected, with default, unauthenticated installations reachable from untrusted networks at greatest risk. JFrog has published technical details and a PoC, EPSS assigns a 66.3% probability of exploitation within 30 days (99th percentile), and news coverage reports active exploitation with 12,000+ instances exposed on the internet. Do: Upgrade Flowise to version 3.0.1 or later. Until upgraded, restrict network access to the instance (firewall rules or an authenticating reverse proxy) and explicitly enable authentication, since default installations before 3.0.1 run without it. Review internet-exposed instances for signs of compromise, such as unexpected npx launches or child OS processes spawned by the Flowise service. | 9.8 | 66% | PoC |
| large≈12,000+ internet-exposed Flowise instances |
Full article560 words · extracted from securityaffairs.com · click to collapse

Attackers are exploiting a critical Flowise flaw, tracked as CVE-2025-59528 (CVSS score of 10), that lets them run malicious code and access systems due to poor validation of user-supplied JavaScript.
Attackers are actively exploiting a critical vulnerability in Flowise, tracked as CVE-2025-59528, that allows remote code execution and file system access. The flaw stems from improper validation of user-supplied JavaScript in a configuration function, exposing systems to full compromise.
Flowise is an open-source platform that lets users build and manage customized LLM (large language model) workflows and autonomous agents. It provides a drag-and-drop interface to design AI flows, connect models, and integrate external tools or APIs without deep programming knowledge. Essentially, it simplifies creating AI-driven applications and automated processes.
The CustomMCP node in Flowise lets users configure connections to external MCP servers, but it processes the mcpServerConfig input insecurely. Instead of validating it, the system executes it as JavaScript. The convertToValidJSONString function passes user input directly to the Function() constructor, running it with full Node.js privileges. This allows access to sensitive modules like child_process and fs, enabling command execution and file system access, making the flaw highly dangerous.
“The CustomMCP node allows users to input configuration settings for connecting to an external MCP (Model Context Protocol) server. This node parses the user-provided mcpServerConfig string to build the MCP server configuration. However, during this process, it executes JavaScript code without any security validation.” reads the advisory. “Specifically, inside the convertToValidJSONString function, user input is directly passed to the Function() constructor, which evaluates and executes the input as JavaScript code. Since this runs with full Node.js runtime privileges, it can access dangerous modules such as child_process and fs.”
The vulnerability lets attackers run arbitrary JavaScript on the Flowise server, leading to full system takeover, file access, command execution, and data theft. Since exploitation only requires an API token, it poses a severe risk to business operations and sensitive customer data.
The flaw impacts Flowise versions up to 3.0.5 and was fixed in version 3.0.6, released in September 2025.
VulnCheck detected first exploitation of CVE-2025-59528, the activity appears to come from a single Starlink IP, with 12,000–15,000 exposed instances online.
“New hashtag#KEV: Early this morning, VulnCheck‘s Canary network began detecting first-time exploitation of CVE-2025-59528, a CVSS-10 arbitrary JavaScript code injection vulnerability in Flowise, an open-source AI development platform. The vulnerability resides in the CustomMCP server logic in multiple versions of Flowise and allows for code execution.” Caitlin Condon, VP, Security Research at VulnCheck wrote on LinkedIn. “Observed activity so far originates from a single Starlink IP. Our team’s ASM queries show 12,000 – 15,000 instances of Flowise on the public internet as of today.”
CVE-2025-59528 is the third Flowise flaw actively exploited in the wild, following CVE-2025-8943 (CVSS score: 9.8) and CVE-2025-26319 (CVSS score: 8.9).
“This is a critical-severity bug in a popular AI platform used by a number of large corporations. This specific vulnerability has been public for more than six months, which means defenders have had time to prioritize and patch the vulnerability,” added Condon.” “The internet-facing attack surface area of 12,000+ exposed instances makes the active scanning and exploitation attempts we’re seeing more serious, as it means attackers have plenty of targets to opportunistically reconnoiter and exploit,” Condon said.
Follow me on Twitter: @securityaffairs and Facebook and Mastodon
(SecurityAffairs – hacking, Flowise)
Text extracted automatically; images, tables and formatting may be missing. Original: https://securityaffairs.com/190471/security/attackers-exploit-critical-flowise-flaw-cve-2025-59528-for-remote-code-execution.html