Executive brief
MCP Calculate Server is a mathematical calculation service used by AI models to perform complex computations. A security flaw allows attackers to bypass AI safety filters and execute malicious commands on the underlying server. This could lead to a total system takeover, unauthorized data access, or service disruption.
Technical details
The MCP Calculate Server (server.py) is vulnerable to remote code execution (RCE) due to the unsafe use of the Python eval() function when processing mathematical expressions. While the service is intended to be used by Large Language Models (LLMs) to evaluate SymPy expressions, an attacker can use prompt injection or direct malicious input to bypass LLM-level restrictions. By crafting an expression that includes Python's __import__ function, an attacker can execute arbitrary system commands with the privileges of the application. The vulnerability is exploited over the network without authentication or user interaction. A fix is available in version 0.1.1 which implements better input handling.
Affected products
- 611711Dark MCP Calculate Server < 0.1.1
Timeline
- 2026-05-04: advisory: GitHub advisory GHSA-2mgq-7rfg-rwpj published
- 2026-05-15: disclosed: CVE-2026-44717 published to NVD