GoKawiilZ.ai's ZCode desktop app secretly uploads users' full Git history to Alibaba Cloud
A reverse-engineering report by developer ferstar found that ZCode, the coding agent built by Z.ai around its GLM models, silently packages a user's entire workspace—including .git history, LFS caches, reflogs and global configs—encrypts it, and uploads the archive to Aliyun OSS whenever a user is logged in. The researcher captured a 313MB encrypted archive from a 345MB commercial project spanning over 42,000 files, plus 564 logged failed upload attempts during investigation.
GoKawiil's interpretation of the reporting above, not reported fact.
The case highlights that open-weight AI models like GLM don't guarantee an open or trustworthy runtime; ZCode's closed harness gave it unchecked access to sensitive source code and credentials without clear consent. Because the encryption uses server-issued public keys, Z.ai—or anyone controlling its infrastructure—could potentially decrypt uploaded data, raising serious concerns for developers using the tool on proprietary codebases.
- ZCode silently exfiltrates entire Git repositories, including history and configs, to Alibaba Cloud storage.
- Open GLM model weights do not mean the ZCode application itself is open source or safe to trust.
- Security researchers are urging developers to avoid closed-source AI coding agents in favor of auditable, open alternatives.
Source: tokenstead.ai, 2026-09-18
Published there as: “ZCode, the GLM coding agent, silently uploads your Git history”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.