btw

Launch a background sub-agent to research side questions via code, shell, and web.

15|5|Updated May 17, 2026
One-click install
npx skills add https://github.com/cropsgg/skills --skill btw-cropsgg
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: btw
Source: https://github.com/cropsgg/skills/tree/main/skills/productivity/btw
Command: npx skills add https://github.com/cropsgg/skills --skill btw-cropsgg

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Side-channel Q&A isolates side questions from the main task by launching a background sub-agent that can search the codebase, run shell commands, or look up documentation when needed, returning only the raw answer to the user.

Core Features & Use Cases

  • Background sub-agent with read, shell, and web access to fetch information on demand
  • Returns the final answer only, without tool transcripts
  • Triggers on /btw or when a side-question appears during an active task
  • Keeps the main thread free from tool traces and intermediate reasoning
  • Supports seeding with @ mentions to seed initial context and targets

Quick Start

Invoke /btw with your side-question and optional seed paths so the sub-agent can perform reads, shell actions, and web lookups and return a single raw answer.

Frequently Asked Questions about btw

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I ask a side question during an active coding task without cluttering the main thread?

A side-channel Q&A isolates side questions by launching a background sub-agent that fetches information and returns only the raw answer, keeping the main workflow free from intermediate tool traces.

Can I run shell commands and look up documentation in the background while working?

Yes, a background worker can execute shell commands and look up documentation on demand. It uses read, shell, and web access to gather information, then suppresses intermediate steps to return only the final answer.

What is the best way to research a side question from code or web sources without interrupting my primary workflow?

Use a background sub-agent to perform side-question research across code, shell, or web sources. It enforces a separate worker that handles the lookups and delivers a single raw answer.

How do I seed initial context when triggering a background Q&A sub-agent?

You can seed initial context and targets for the sub-agent by using @ mentions alongside your side question, allowing the background worker to perform accurate reads, shell actions, and web lookups.

Does the background Q&A worker return tool transcripts or intermediate reasoning steps?

No, the background worker suppresses intermediate steps and tool transcripts entirely. It enforces a clean response that returns only the final raw answer to the user.

When should I avoid using a side-channel Q&A for production engineering tasks?

Avoid side-channel Q&A if you need to review intermediate tool traces or debug the sub-agent's reasoning steps, because this approach strictly suppresses all intermediate output in favor of a single raw answer.