CVE-2026-65698 is a path traversal vulnerability in void editor (void) through version 1.3.4. The vulnerability exists in the AI agent's file-reading tools, which accept absolute paths or file:// URIs without restricting access to the open workspace directory. Attackers can inject instructions that cause these tools to read arbitrary files on the host system, bypassing any approval gates designed to prevent unauthorized access. This can lead to unauthorized disclosure of sensitive files, including SSH private keys and cloud credentials. The vulnerability has a CVSS 4.0 score of 6.0, indicating medium severity, with network attack vector, high complexity, no privileges required, and partial user interaction needed.

An attacker who is network-adjacent can exploit this vulnerability to read arbitrary files on the host system outside the intended workspace. This can result in the disclosure of sensitive information such as SSH private keys and cloud credentials. The vulnerability bypasses approval gates, allowing silent exfiltration of data. There is no indication of privilege escalation or code execution. No known exploits in the wild have been reported.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Until a fix is available, users should restrict network access to the void editor AI agent tools and avoid processing untrusted content that could inject malicious instructions. Monitor vendor communications for official patches or updates addressing this vulnerability.