🧪 Test Script Runner

Another one from the pack's "Utils🛠" trio that isn't about image generation at all - it's a remote control. One dropdown, two positions: START or STOP. Going purely by the name and that binary action set, this reads as exactly what it says on the label: a node the author built to kick off, and later kill, some external test script from inside a ComfyUI graph rather than switching to a terminal window.

Same caveat that applies to its two siblings in this trio (xO_ComfyUIPortRunner and xO_WorkflowRunner): there's no description shipped with this node and the README doesn't mention it at all - the README covers exactly two of the pack's eight nodes, both Ollama-related, and this isn't one of them. What follows is a reading of the schema and the category it's filed under, not a look at the source.

How it works

There's genuinely not much surface area here: action picks a mode, the node presumably starts or terminates whatever script it's wired to run, and reports back a status string. No further configuration is exposed - no script path, no arguments - which suggests either a hardcoded target in the node's own code, or a companion piece of setup this README doesn't cover.

The inputs and outputs that matter

• action - an enum, START or STOP, defaulting to START. The only input.
• The output - labeled plainly STRING in the schema - most likely reports what happened: started, already running, stopped, not found. The usual shape for a control node like this.

How to install it

Standard pack install. ComfyUI Manager: search ComfyUI Neural Nodes, install, restart. Manually:

cd ComfyUI/custom_nodes
git clone https://github.com/xobiomesh/ComfyUI_xObiomesh

then restart. No models, nothing heavy - this is control-flow, not generation.

Common issues & troubleshooting

There's no crowd-sourced list of gotchas to lean on here - this pack has essentially no public footprint, and this specific node doubly so. A few honest notes instead:

This is the one node in the whole pack worth pausing on before you use it, specifically because of what its job is. Custom nodes already run as plain Python with your full OS permissions the instant ComfyUI loads them - there's no sandboxing anywhere in this ecosystem, which is the exact mechanism a real malicious node once used against unsuspecting users through completely normal install channels. A node whose entire purpose is starting and stopping an external script is that same lack of sandboxing aimed at something explicit rather than incidental. That's not a claim about bad intent here - everything about this trio reads like a personal dev-testing convenience the author happened to ship alongside the rest of the pack - it's just the honest framing: know what script this actually points at on your machine before you set action to START.

If START seems to do nothing visible, wire the output STRING into this pack's own ShowText_xO node before assuming it failed silently - a status message not being displayed anywhere is a much more common cause than the action itself not firing.

Most workflows have no reason to include this node at all. Unless you're specifically automating your own workflow-testing loop the way the pack's dev-tool trio implies the author was, it's not a general-purpose utility - skip it.