Purpose
This tool acts as a bridge connecting the Claude Code CLI to Feishu (Lark) and email, allowing users to manage autonomous digital employees, monitor ongoing AI sessions, and approve permissions via a chat interface.

Security Assessment
The overall security risk is rated as High. The scan detected multiple instances of synchronous shell command execution (`execSync`) within the codebase, specifically in `daemon.ts` and `project-lock.ts`. If an attacker successfully injects a malicious prompt or compromises an incoming message, these synchronous executions could allow arbitrary command execution on the host server. Additionally, the tool requires outbound network requests, file system access, and environment variable access to facilitate its core messaging and monitoring functions. While no hardcoded secrets or explicitly dangerous permissions were found, the execution flaws pose a significant threat.

Quality Assessment
The project is very new and currently has low community visibility with only 9 GitHub stars, meaning it has not been broadly tested or vetted by a wide audience. However, the repository is under active development (last push was today), and the code is legally unencumbered under the standard MIT license. The README is comprehensive and clearly documents the tool's constraints, noting that the underlying Claude channel feature is not yet fully public.

Verdict
Use with caution — the tool is highly useful for remote AI management but poses significant security risks due to its use of `execSync`.