Hackers Exploit Windmill Flaw to Read Arbitrary Server Files Without Authentication

A high-severity security flaw impacting open-source developer platform Windmill has come under active exploitation in the wild, per VulnCheck. The vulnerability in question is CVE-2026-29059 (CVSS score: 7.5), a case of unauthenticated path traversal impacting Windmill's "get_log_file" endpoint ("/api/w/{workspace}/jobs_u/get_log_file/{filename}"). "The filename parameter is concatenated into

The article reports an actively exploited unauthenticated path traversal vulnerability in Windmill (CVE-2026-29059), where the get_log_file endpoint concatenates the filename parameter into a file path without sanitization, allowing attackers to read arbitrary server files using ../ sequences.[1][7] Reported impacts include exposure of sensitive environment variables such as SUPERADMIN_SECRET via /proc/1/environ, which can be used as a bearer token for superadmin access and arbitrary code execution through Windmill’s job preview API, though SUPERADMIN_SECRET is not set by default.[1][9][10] From a RealGround perspective, any AI or automation workloads running on Windmill or integrated platforms risk compromise of API keys, model credentials, and orchestration tokens, enabling attackers to hijack AI agents, alter workflows, or exfiltrate model-access secrets. Organizations should harden AI-related deployments by upgrading to fixed versions, isolating Windmill from core AI infrastructure, and continuously red-teaming agent workflows to detect abuse paths from arbitrary file read to AI-agent-driven RCE.