CVE-2026-61445 is a critical vulnerability in MervinPraison's PraisonAI product affecting versions prior to 4.6.78. The vulnerability stems from improper limitation of pathnames and lack of command sanitization in the AICoder component's handling of LLM tool calls. This allows attackers to inject malicious prompts through the chat interface, resulting in arbitrary file write and command execution with root privileges. The CVSS 4.0 base score is 9.4, reflecting network attack vector, low complexity, no user interaction, and high impact on confidentiality, integrity, availability, and security requirements. No patch or official remediation level has been disclosed as of the publication date.

Successful exploitation enables an attacker to write arbitrary files to any location on the filesystem and execute arbitrary shell commands with root privileges, leading to full system compromise. This impacts confidentiality, integrity, and availability of the affected system. The vulnerability can be triggered remotely over the network without user interaction and requires only low privileges to exploit.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Until an official fix is released, restrict access to the chat interface and monitor for suspicious input patterns that may attempt to exploit this vulnerability. Avoid exposing the vulnerable AICoder component to untrusted users.