
rimz
worthwhileA realtime mult-agent fleet dashboard that brings observability and control to parallel agentic coding inside tmux/Zellij.



What it is
A Rust binary that runs inside a tmux or Zellij session, providing a realtime sidebar dashboard showing state, cost, and context health for multiple coding agents (Claude, Codex, etc.), plus fleet-level control via messages, worktrees, and scheduled loops.
How it differs from vanilla Claude
Vanilla Claude runs one instance at a time in your terminal; Rimz orchestrates dozens, shows them all at once with live token costs, and lets you message them by handle and route attention interactively.
Skill, plugin, or workflow shift?
standalone-app: Rimz is a separate binary that manages agent panes and the dashboard; it doesn't hook into an existing IDE or editor but replaces no prior tool altogether.
Devil's advocate — is this just complexity?
You're adding a whole layer of complexity—another binary, tmux/Zellij dependency, per-agent hook config—just to watch agents work in parallel. For most workflows, running one Claude in a terminal, maybe two in separate tmux panes, and using Claude's own 'multiplex' is enough. The dashboard is pretty but adds marginal value over a simple tmux layout. If you only run 1-2 agents, this is overkill. Also, the project is beta and has barely any traction (22 stars); it could vanish. The auto-continue and mobile steering are thin polish over what a simple script (e.g., retry on rate limit) already does. You're paying setup cost for a problem you likely don't have at scale.
What would make it better
Add pre-built agent templates that auto-configure hooks for the most common agents (Claude, Codex, OpenCode) without manual toml editing. Provide a quick-start 'demo' mode that runs a mock fleet so users see value before wiring up real agents. Ship official integrations for VS Code and JetBrains so the dashboard appears as a panel, not only inside tmux. Finally, allow running without tmux/Zellij (e.g., a standalone TUI) to widen adoption.
The honest case for it
If you routinely run 10+ agents in parallel across multiple providers, Rimz is the only tool that gives you a single pane of glass for all their state, cost, and attention routing. The worktree isolation and team compositions are genuinely useful for large feature branches with separate planning/coding/review agents. The scriptable interface with exit codes and JSON output makes it automation-friendly for CI. For power users of tmux/Zellij who already think in multiplexers, Rimz adds real leverage with minimal friction.
Who it's for
Audience fit
Depth and leverage for a technical engineer who wants to understand it and level up their workflow — not just offload work.
Value for someone who wants a more capable tool without the technical depth — accessible, does-it-for-you.
Deep orchestration for engineers running multiple agents in parallel; vibe coders will find the setup complex and configuration-heavy.