Does your coding agent get stuck in loops or claim work is done when it isn’t?

A developer is seeking 3-5 developers to test a new reliability monitoring tool for AI coding agents. The tool aims to detect when agents like Claude Code and Codex get stuck in loops (e.g., fix → test → fail → fix) or claim tasks are complete when they are not. The developer is looking for honest feedback on the tool's effectiveness in catching problems during real coding sessions.