CVE-2026-19282 describes a command injection vulnerability in the andreahaku llm_memory_mcp component, specifically in the auto.capture function of src/autolearn/GitHooksManager.ts. An attacker with local access can manipulate the 'hash' argument to execute arbitrary commands. The product lacks versioning, so no specific affected versions are known. The vendor has not responded to the issue report, and no official remediation or patch is available. The CVSS 4.0 vector indicates local attack vector, low attack complexity, no user interaction, and partial impacts on confidentiality, integrity, and availability.

Successful exploitation allows an attacker with local access and low privileges to execute arbitrary commands on the affected system, potentially compromising confidentiality, integrity, and availability to a limited extent. Remote exploitation is not possible according to the data. The medium severity reflects the limited scope and required local access.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Since the vendor has not responded and no patch is available, users should restrict local access to trusted users only and monitor for suspicious activity related to the vulnerable function. Avoid passing untrusted input to the 'hash' argument in the auto.capture function if possible.