
agent-setup
How I run Claude Code across ~90 repos as a solo builder: rules, sub-agents, workflows, lessons
Install with your AI
Paste into Claude Code, Cursor, or any agent — it reads the repo and wires the tool into your project.
Install and set up agent-setup (git-clone project) into my current project. Found on https://claudeers.com/agent-setup Repo: https://github.com/lisbeth718/agent-setup Homepage/docs: — Detected install method: git-clone → git clone https://github.com/lisbeth718/agent-setup Category: automation. Platforms: api, web. Read the repo's README for exact setup and env vars, then install it and wire it into my project. Claudeers Health Verdict: active; community-verified: false. Confirm the source before running anything.
git clone https://github.com/lisbeth718/agent-setup
// compatibility
| Platforms | api, web |
|---|---|
| Operating systems | — |
| AI compatibility | claude |
| License | MIT |
| Pricing | open-source |
| Language | — |
agent-setup
The global rules, sub-agents and skills I use, plus slash commands and a settings template that package the same loops, to run Claude Code and other coding agents across about 90 repos as a solo builder. Deploys are push-to-deploy on my own servers. For anyone who works alone or in a small team and wants a lean, opinionated setup to copy from.
Core ideas
- Role-pinned models: planning on the main session, build work on Sonnet fixers, review on Opus hunters. Every sub-agent is a pinned type, nothing inherits.
- The reviewer is never the builder. Reviews run in a clean context.
- Verified vs reasoned: agents must say what they ran and what they only read.
- Brief output: answer first, only what changes the next action.
- Push-to-deploy, no CI: push the default branch, the build server builds. Keep lint and types clean instead.
Repo map
AGENTS.md: tool-agnostic rules (model routing, folder layout, deploy).CLAUDE.md: global Claude Code rules (output style, sub-agent models).agents/:fixer(scoped edits, Sonnet),hunter(read-only review, Opus),researcher(web research, Opus).skills/brief/: output-length mode.commands/: slash commandstriage-alert,ship,review,prepublish,sweep.settings/: examplesettings.json(auto mode, model effort) and notes.WORKFLOW.md: the four loops and prompt templates.LESSONS.md: each rule and what went wrong before it existed.
Install
mkdir -p ~/.claude/agents ~/.claude/skills ~/.claude/commands ~/Development
cp -n agents/*.md ~/.claude/agents/
cp -rn skills/* ~/.claude/skills/
cp -n commands/*.md ~/.claude/commands/
cp -n CLAUDE.md ~/.claude/CLAUDE.md # -n never overwrites; if you already have one, append by hand
cp -n AGENTS.md ~/Development/AGENTS.md # or your projects root
settings/settings.example.json is not copied: merge it into ~/.claude/settings.json by hand (see settings/README.md). Edit model IDs to whatever you have access to and adjust the folder layout to your own.
Read next: WORKFLOW.md, LESSONS.md.
License
MIT
// faq
What is agent-setup?
How I run Claude Code across ~90 repos as a solo builder: rules, sub-agents, workflows, lessons. It is open-source on GitHub.
Is agent-setup free to use?
agent-setup is open-source under the MIT license, so it is free to use.
What category does agent-setup belong to?
agent-setup is listed under automation in the Claudeers registry of Claude-compatible tools.
// embed badge
[](https://claudeers.com/agent-setup)
// retro hit counter
[](https://claudeers.com/agent-setup)
// reviews
// guestbook
// related in Automation & Workflows
The API to search, scrape, and interact with the web at scale. 🔥
🌐 Make websites accessible for AI agents. Automate tasks online with ease.
Taste-Skill - gives your AI good taste. stops the AI from generating boring, generic slop