See a running agent command in full: command line, live output, exit status (incl. the waiting strip)

[JochenVanSeveren]

The problem

When an agent babysits a deploy (merge, wait for the build, wait for the host's deploy), it runs long watch commands. In the thread each one is a single truncated row, e.g. Command "osh builds watch …, or the waiting strip Waiting · vercel inspect … with ⓘ and Stop.

From there I can't see:

• the full command line (all arguments, cwd)
• the live stdout/stderr while it runs
• elapsed time, exit code, or whether it's healthy or hung

So I either trust the agent's summary or hit Stop on the whole turn.

Proposal

An Open Terminal action on the command row and on the waiting strip, showing:

1. the full command and cwd, with a copy button
2. streaming output with scrollback, still available after the command ends
3. status, elapsed time and exit code
4. a Stop for that command only, where the provider allows it

View + interrupt is enough for a first version.

Environment

macOS desktop app, Claude provider (Opus 5.5). Screenshot below, identifying details redacted.

Related

• Open Terminal from an agent command: agent-owned tabs grouped by thread / lineage #30557 Open Terminal from an agent command (closest; this is a concrete case for it, including the waiting strip)
• Show a background task's progress in the waiting strip while an agent waits on it #31460 progress / latest output line in the waiting strip
• Let agents use a real terminal in the thread that I can watch and type into #29136 shared agent terminal