omp packages

@howaboua/pi-smart-btw0.2.8

Async side-session questions for Pi with explicit injection into the main chat.

howaboua Published 27 days ago MIT ~998 downloads/mo

Untested on omp

This is a pi package. No one has reported whether it works on omp yet. If you try it, add your result below.

Did it work for you?

Result

One report per person. A new report replaces your old one.

Reports 0

No reports yet.

Readme

@howaboua/pi-smart-btw

Adds /btw for questions you do not want to derail the main chat. Each numbered slot runs an async, ephemeral child Pi RPC process. Its answers stay out of the main model context until you inject them.

Install

pi install npm:@howaboua/pi-smart-btw

Usage

/btw 1 what is this repo?
/btw 2 explain this error
/btw 1 continue that answer
/btw
  • /btw N <question> sends to slot N.
  • /btw <question> sends to the active slot.
  • /btw N switches slots; /btw opens the panel.
  • Each slot has its own queue and child, so a slow answer in one does not block another.

Answers live in the transcript and survive restarts, but stay outside the main model context until injected. The widget only shows status and controls.

Controls

  • Alt+C — inject the active slot's answers into the main chat, then clear it
  • Alt+X — clear the active slot without injection
  • Alt+Z — prefill /btw
  • Alt+H / Alt+L — previous/next slot
  • Alt+1 … Alt+9 — jump to a slot
  • Alt+J / Alt+K — fold/unfold the widget
  • /btw config — model, thinking, shortcut, and link settings

Configuration

Settings live at ~/.pi/agent/pi-smart-btw.json. New configurations default to openai-codex/gpt-6-luna with low thinking. /btw config edits provider, model, and thinking. Edit shortcuts opens the JSON file in $VISUAL or $EDITOR; reload after changing shortcuts or advanced options such as command. Thinking is clamped to the selected model.