Use Herdr when an operator needs to coordinate several terminal-based coding agents in the same project without losing track of blocked, working, and completed sessions. The workflow is to start Herdr in the repository or workspace, launch agent CLIs in separate panes, watch real terminal output, detach while work continues, reattach later, and use the documented session state and keyboard controls to review or hand off work. This is especially useful for parallel implementation, long-running coding-agent jobs, and remote terminal work where the agent process should keep running after the operator steps away. Invoke this instead of running a single agent CLI directly when the job needs visible multi-agent supervision, persistent terminal sessions, and repeatable operator control over panes and state. The scope boundary is terminal coordination for coding-agent workflows. It is not a general IDE listing, a generic tmux card, or a claim that Herdr supplies the model or writes the code by itself.