Running Commands

Running Commands

Running executes a Snip's resolved command in its configured shell, inside an embedded terminal you can watch and type into — and records the run to your searchable Run History.

Run a command

1. On a Snip card, click Run.
2. Fill in the parameters (the same flyout as Copy).
3. Review the dry-run preview, then click Run.

The dry-run preview

Because executing an arbitrary command is a footgun, Run always shows a preview first — even for a Snip with no parameters. The preview shows exactly what's about to happen:

• the resolved command,
• the shell it will run in (e.g. "Windows PowerShell"),
• the working directory it will run in.

Click Run to proceed, or cancel to back out. Nothing executes until you confirm.

The live terminal

The command runs under a real pseudo-terminal, so it behaves exactly as it would in a normal console:

• Full fidelity — colours, spinners, progress bars and interactive prompts (such as selection or confirmation prompts) all render and work.
• You can type. Keystrokes go to the program, so interactive prompts are usable.
  • Ctrl+C sends an interrupt to the running command.
  • Ctrl+Shift+C (or Ctrl+C while text is selected) copies the selection.
• Cancel stops the run and kills the whole process tree.
• A status line shows a spinner and elapsed time while running, then an exit-code badge and duration when the command finishes.
• When finished you can Copy command or Run again.

Run lifecycle

stateDiagram-v2
    [*] --> Pending: Run confirmed
    Pending --> Running: process starts
    Running --> Finished: process exits
    Running --> Cancelled: you press Cancel
    Finished --> [*]: recorded to history
    Cancelled --> [*]: recorded to history

Loading

Whatever the outcome, the run is saved to Run History with a clean transcript and the retained raw output for colour replay.

Which shell runs the command

A run uses the shell configured on the Snip's CLI (Windows PowerShell by default), unless the Snip overrides it. The options are Command Prompt, Windows PowerShell, PowerShell (pwsh), Bash, or a Custom shell — see Managing CLIs for how each is launched and how to set an executable path and working directory. A Snip can override the CLI's shell and working directory via its Advanced overrides (see Creating and Editing Snips).

Validation happens at run time

If the CLI has an executable path set and it's missing when you run, Snipdeck shows a friendly notice ("couldn't find <path> — has it been installed? edit the CLI to fix the path") instead of failing obscurely. If a Custom shell has no path set, you're told to set it. This is why these fields aren't validated when you save a CLI — you can author Snips for tools you haven't installed yet, and only hit the check when you actually run.

What gets recorded

When a run ends, Snipdeck stores its command, start time, duration and exit code, a clean plain-text transcript (no escape codes, no cursor artefacts, no stacked spinner frames), and the raw output stream (compressed) so the run can be replayed in full colour later. Output is capped (5 MB by default) and truncated beyond that. See Run History and Settings.