explaingit

tuchsoft/zed-use-subagent-skill

Analysis updated 2026-05-18

0JavaScriptAudience · developerComplexity · 2/5Setup · moderate

TLDR

A temporary Zed editor skill that lets you assign different AI models to different subagent task types, like coding versus exploring, via a small command line tool.

Mindmap

mindmap
  root((repo))
    What it does
      Sets subagent model per task type
      zsetagent CLI tool
      Regenerates SKILL.md
    Tech stack
      JavaScript
      Zed editor skill
      Unix symlinks
    Use cases
      Pick model for coding tasks
      Pick model for exploring tasks
      Write subagent instructions
    Audience
      Zed editor developers

Code map

Detail Auto

An interactive map of this repo's files and how they connect — its source is parsed live in your browser. Click Visualize to build it.

filefunction / class

What do people build with it?

USE CASE 1

Assign a stronger, slower AI model to coding subagent tasks in Zed.

USE CASE 2

Assign a faster, cheaper model to codebase exploration or browser testing subagent tasks.

USE CASE 3

Write reusable instruction snippets that get included whenever a task is delegated to a given subagent type.

USE CASE 4

Regenerate the skill's own documentation file after changing your model configuration.

What is it built with?

JavaScriptZed

How does it compare?

tuchsoft/zed-use-subagent-skill00kaku/gallery-slider-block0xkinno/vellum
Stars00
LanguageJavaScriptJavaScriptJavaScript
Last pushed2021-05-19
MaintenanceDormant
Setup difficultymoderateeasyhard
Complexity2/52/54/5
Audiencedevelopergeneraldeveloper

Figures from each repo's GitHub metadata at analysis time.

How do you get it running?

Difficulty · moderate Time to first run · 30min

Requires manual symlinking into PATH and does not work on Windows.

In plain English

This project is a skill for the Zed code editor that helps you control which AI model handles which kind of task when Zed delegates work to a subagent. The author calls it a temporary workaround, meant to be dropped once Zed ships its own built in way to manage subagents, and the project will stop being maintained once that happens. Once installed, it adds a command line tool called zsetagent. With it you can list the task types you have configured, such as coding, browser testing, or exploring a codebase, and assign a specific AI model and provider to each one. For example you might want a stronger, slower model for writing code but a faster, cheaper model for simply exploring files. Each task type can also have its own instruction file describing what that kind of subagent is good at, when to use it, and a required snippet of text to always include when handing it work. Setup involves cloning the project into a specific folder used by Zed's agent skills, then symlinking the zsetagent script into a folder already on your command line path, editing a configuration file with your preferred models, and finally running a command that regenerates the skill's documentation file from that configuration. The README stresses that none of this should happen automatically. The tool is meant to walk you through the setup step by step and ask for confirmation before writing or downloading anything. The project explicitly does not support Windows, because it relies on Unix style file paths and symbolic links. It is written in JavaScript and is aimed at developers already using Zed and its subagent delegation feature who want more control over model choice than the editor currently offers on its own. The README does not state a license for this project.

Copy-paste prompts

Prompt 1
Help me configure zsetagent so coding tasks use a stronger model and exploring tasks use a faster one.
Prompt 2
Walk me through installing use-subagent into my Zed agents skills folder and symlinking zsetagent to my PATH.
Prompt 3
Write a config/coding.md instruction file for this skill describing when to delegate to the coding subagent.
Prompt 4
Explain what zsetagent --skill --force does before I run it.

Frequently asked questions

What is zed-use-subagent-skill?

A temporary Zed editor skill that lets you assign different AI models to different subagent task types, like coding versus exploring, via a small command line tool.

What language is zed-use-subagent-skill written in?

Mainly JavaScript. The stack also includes JavaScript, Zed.

How hard is zed-use-subagent-skill to set up?

Setup difficulty is rated moderate, with roughly 30min to a first successful run.

Who is zed-use-subagent-skill for?

Mainly developer.

Open on GitHub → Explain another repo

This repo across BitVibe Labs

Verify against the repo before relying on details.