CVE-2026-7216

Vulnerability

Overview

The create_sketch tool in processing_server.py concatenates the user-supplied sketch_name directly into filesystem paths using os.path.join() without canonicalization or base-directory enforcement [1][2]. This allows an attacker to inject path traversal sequences such as ..\..\Desktop\evil to escape the intended PROC:\Users\chelo\OneDrive\Documentos\Processing` directory [1].

Exploitation

No authentication is required; the MCP server listens for commands from Claude Desktop, and any client that can invoke the create_sketch tool can supply a malicious sketch_name [1][2]. The server then creates directories and writes .pde files at the attacker-controlled path, enabling arbitrary file creation or overwrite outside the intended Processing sketch root [1].

Impact

An attacker can write arbitrary .pde files to arbitrary locations on the host filesystem, potentially overwriting existing files or planting malicious sketches that could be executed later [1]. The vulnerability is remotely exploitable and proof-of-concept code has been publicly released [1].

Mitigation

As of the report date (April 10, 2026), no fix has been provided by the vendor [1]. The project follows a rolling release model and has not responded to the issue report [1]. Users should restrict access to the MCP server or implement input validation on sketch_name to block traversal sequences until an official patch is available.