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

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.

Three vulnerabilities in the mcp-server-git, maintained by Anthropic, allow file access, deletion, and code execution via prompt injection. The flaws have been addressed in versions 2025.9.25 and 2025.12.18. The vulnerabilities include path traversal and argument injection issues that can be exploited to manipulate Git repositories and execute arbitrary code. The issues were disclosed by Cyata researcher Yarden Porat, highlighting the risks of prompt injection attacks without direct system access. The vulnerabilities affect all versions of mcp-server-git released before December 8, 2025, and apply to default installations. An attacker only needs to influence what an AI assistant reads to trigger the vulnerabilities. The flaws allow attackers to execute code, delete arbitrary files, and load arbitrary files into a large language model's context. While the vulnerabilities do not directly exfiltrate data, sensitive files may still be exposed to the AI, creating downstream security and privacy risks. The vulnerabilities have been assigned CVE-2025-68143, CVE-2025-68144, and CVE-2025-68145.

The **U.S. Cybersecurity and Infrastructure Security Agency (CISA)** has added **CVE-2025-68613** to its **Known Exploited Vulnerabilities (KEV) catalog**, mandating federal agencies to patch n8n instances by **March 25, 2026**, due to **active exploitation** of this critical remote code execution (RCE) flaw. Meanwhile, **Pillar Security** has disclosed two new critical vulnerabilities (**CVE-2026-27577** and **CVE-2026-27493**), with the latter being a **zero-click, unauthenticated flaw** that allows **full server compromise** via public form endpoints (e.g., a "Contact Us" form) without requiring authentication or user interaction. Over **40,000 unpatched instances** remain exposed globally, with **18,000+ in North America and 14,000+ in Europe**, per Shadowserver data. This development follows a series of **critical n8n vulnerabilities** disclosed since late 2025, including **CVE-2026-21877 (CVSS 10.0)**, **CVE-2026-21858 (unauthenticated RCE)**, and **four March 2026 flaws (CVE-2026-27577, CVE-2026-27493, CVE-2026-27495, CVE-2026-27497)** enabling **sandbox escapes, credential theft, and unauthenticated expression injection**. Affected versions span **<1.123.22, >=2.0.0 <2.9.3, and >=2.10.0 <2.10.1**, with patches available in **1.123.22, 2.9.3, and 2.10.1**. The platform’s widespread use in **AI orchestration and enterprise automation**—coupled with its storage of **API keys, database credentials, and cloud secrets**—makes it a prime target for attackers seeking **full server compromise** or **lateral movement into connected systems**.

Active exploitation of critical vulnerabilities in Gladinet's CentreStack and Triofox products continues. The zero-day vulnerability, CVE-2025-11371, is an unauthenticated local file inclusion bug that allows unintended disclosure of system files. This flaw affects all versions prior to and including 16.7.10368.56560. The vulnerability has been exploited to retrieve the machine key from the application Web.config file, enabling remote code execution via a ViewState deserialization vulnerability. Three customers have been impacted so far. A patch for the zero-day vulnerability CVE-2025-11371 is now available in CentreStack version 16.10.10408.56683. Users are advised to upgrade to this version or, if upgrading is not possible, disable the "temp" handler within the Web.config file for UploadDownloadProxy to mitigate the risk. The vendor, Gladinet, has been notified and is working on a fix. The vulnerability was detected by researchers at Huntress on September 27, 2025. The flaw was exploited to obtain a machine key and execute code remotely. The attack used an older deserialization vulnerability (CVE-2025-30406) to achieve remote code execution (RCE) through ViewState. The mitigations will impact some functionality of the platform but prevent exploitation of CVE-2025-11371. The U.S. Cybersecurity and Infrastructure Security Agency (CISA) added CVE-2025-11371 to its Known Exploited Vulnerabilities (KEV) catalog on November 5, 2025, citing evidence of active exploitation. Federal Civilian Executive Branch (FCEB) agencies are required to apply the necessary fixes by November 25, 2025, to secure their networks. Additionally, a new critical vulnerability, CVE-2025-12480 (CVSS score: 9.8), has been discovered in Gladinet's Triofox file-sharing and remote access platform. This flaw allows attackers to bypass authentication and access configuration pages, resulting in the upload and execution of arbitrary payloads. The threat cluster tracked as UNC6485 has been exploiting this flaw since August 24, 2025. The attackers have used the built-in antivirus feature to execute malicious files and set up encrypted tunnels to command-and-control servers, leveraging remote access tools like Zoho Assist and AnyDesk for further exploitation. The vulnerability CVE-2025-12480 was discovered and reported by Mandiant on November 10. The flaw allows an attacker to gain access to initial setup pages even after setup is complete, enabling the upload and execution of arbitrary payloads. The exploitation campaign started on August 14, 2025. The attackers exploited an HTTP Host header vulnerability by spoofing localhost in requests, bypassing access controls to reach the normally restricted AdminDatabase.aspx setup page. The flaw stemmed from missing origin validation and over-reliance on the host header, allowing unauthenticated remote access to critical configuration pages. The attackers logged in using the newly created Admin account and uploaded malicious files to execute them using the built-in anti-virus feature. A new actively exploited vulnerability in Gladinet's CentreStack and Triofox products has been disclosed, stemming from the use of hard-coded cryptographic keys. This flaw affects nine organizations so far. The use of hard-coded cryptographic keys could allow threat actors to decrypt or forge access tickets, enabling them to access sensitive files like web.config that can be exploited to achieve ViewState deserialization and remote code execution. The attacks involve specially crafted URL requests to the "/storage/filesvr.dn" endpoint, with the Username and Password fields left blank, causing the application to fall back to the IIS Application Pool Identity. The timestamp field in the access ticket is set to 9999, creating a ticket that never expires, allowing threat actors to reuse the URL indefinitely to download the server configuration. Organizations using CentreStack and Triofox are advised to update to the latest version, 16.12.10420.56791, released on December 8, 2025, and scan logs for the presence of the string "vghpI7EToZUDIZDdprSubL3mTZ2," which is the encrypted representation of the web.config file path. In the event of indicators of compromise (IoCs), it is imperative to rotate the machine key by generating new keys in the IIS Manager and restarting IIS after repeating the same step for all worker nodes.

A command injection vulnerability (CVE-2025-53967) in the Figma MCP server allows remote code execution. The flaw, stemming from unsanitized user input, was patched in version 0.6.3. The issue affects developers using AI-powered coding agents like Cursor. The vulnerability could be exploited by attackers on the same network or via DNS rebinding attacks. It was discovered by Imperva in July 2025 and was addressed in the latest release. The flaw resides in the 'src/utils/fetch-with-retry.ts' file, where the curl command is constructed using shell command strings, enabling potential remote code execution. The patch replaces 'child_process.exec()' with 'child_process.execFile()' and implements proper input validation. Users should upgrade to Figma MCP version 0.6.3 or higher, audit systems using vulnerable versions, and review logs for suspicious command execution patterns. There are over 15,000 MCP servers in the world, with many misconfigured and lacking authentication or access controls.