Vision

Running several coding agents at once, the bottleneck stops being the work and becomes attention: which of these twelve tabs is blocked on me, and for what? Rig answers that from the tab rail. Agent status is not a plugin bolted onto a terminal — it is the thing the terminal is built around, so the answer is always on screen instead of one tab-switch away.

The Problem

A terminal treats every tab as an anonymous shell. An agent working for four minutes and an agent that stopped ninety seconds ago to ask permission look identical until you go look. Existing agent-status add-ons flatten every kind of interruption into one ambiguous "needs attention" signal, which trains you to ignore it — the one state that actually blocks progress is indistinguishable from a nudge.

Key Differentiators

Status engine at the core: a pure, framework-agnostic engine ingests agent hook events and owns the status model; the UI is built around it, not the other way round.
Interruptions are not interchangeable: a permission request, a questionnaire, and an idle timeout are distinct states, because they demand different things from you.
Quiet about bursts: agents emit several events for a single prompt; the engine coalesces them so one prompt is one signal.
Session metadata that means something: session id, title, and live CPU/RAM per tab, so a runaway agent is visible before it is expensive.
Small and native: Tauri 2 with a Rust core and the system webview, not a bundled browser.

Rig Changelog