// 6,413 real setups indexed · updated 2026-10-09
See how the best AI-agent setups are actually wired.
AgentRigs indexes public Claude Code, Codex & Cursor configs — MCP servers, skills, subagents, hooks, permissions — and ranks every component by real-world adoption. Find a rig you like, copy it in one command.
New: Rig Doctor. Paste your CLAUDE.md / settings.json / .mcp.json or a repo URL: context tax, guardrail score and fixes. check my rig →
nicknisi/dotfiles
★ 2,999 · Claude Code
★ 2,999 · Claude Code
context tax
28.8k tok
28.8k tok
guardrails
3/5
3/5
archetype
YOLO Cowboy
YOLO Cowboy
$ claude mcp add sessions -- sessions --mcp $ claude mcp add --transport http claude-design https://api.anthropic.com/v1/design/mcp -H 'Authorization: YOUR_VALUE' $ npx degit nicknisi/dotfiles/home/.claude ./rig-dotfiles # inspect, then merge into .claude/
6,413public rigs
4,104shared components
2.2kmedian context tax (tok)
9%block destructive cmds
Most-used MCP servers
Counted once per rig, across 6,413 public setups. Full leaderboard →
| 1 | Context7 | 170 | |
| 2 | Playwright MCP | 113 | |
| 3 | GitHub MCP | 104 | |
| 4 | Sequential Thinking | 67 | |
| 5 | Chrome DevTools MCP | 60 | |
| 6 | Filesystem | 49 | |
| 7 | Memory (knowledge graph) | 43 | |
| 8 | Exa | 35 | |
| 9 | Notion | 35 | |
| 10 | Linear | 33 |
Popular skills & subagents
Same name appearing in multiple independent rigs.
handoff · 215skill-creator · 211code-review · 209frontend-design · 185review · 174grill-me · 173commit · 163tdd · 155
code-reviewer · 290security-reviewer · 126architect · 122planner · 119reviewer · 106security-auditor · 81researcher · 79debugger · 72
What's your archetype?
Top personal setups
all setups →
Yeachan-Heo/oh-my-claudecode
Teams-first Multi-agent orchestration for Claude Code
Orchestrator 14.5k tok ·
lewagon/dotfiles
Default configuration for Le Wagon's students
Minimalist 0 tok ·
tigerless-labs/autoharness
Autoharness — a self-learning skill layer for Claude Code — distills skills from your real sessions, updates them as you work, and prunes the ones that stop getting used. No daemon, no benchmark.
Pragmatist 2.6k tok ·
anthropics/defending-code-reference-harness
Skills for threat modeling, scanning, triage, patching, plus an autonomous scanning harness you can /customize
Pragmatist 4.3k tok ·
paulirish/dotfiles
paul's fish, bash, git, etc config files. good stuff.
Orchestrator 1.3k tok ·
wquguru/harness-books
📚 Two books on harness engineering — the design philosophies behind Claude Code & Codex: constraints, query loops, context governance, multi-agent verification. harness-books.agentway.dev
Minimalist 1.0k tok ·
Chachamaru127/claude-code-harness
Claude Code Dedicated Development Harness - Achieving High-Quality Development Through an Autonomous Plan→Work→Review Cycle
Orchestrator 9.4k tok ·
nicknisi/dotfiles
vim, zsh, git, homebrew, neovim - my whole world
YOLO Cowboy 28.8k tok ·
Recently updated setups
more →
odakin/claude-config
Shared conventions, mechanical enforcement (hooks + scripts), and an operational knowledge base for running many projects and machines with Claude Code
Pragmatist 12.9k tok ·
maolonglong/nix-config
My nix config for macOS (modified from https://github.com/ryan4yin/nix-config)
Minimalist 1.4k tok ·
MithunWijayasiri/dotclaude
My Claude Code setup & Skills.
Skill Collector 1.4k tok ·
Iski08/dotclaude
Multi-agent config for Claude Code with specialized agents and templates to speed code review, refactoring, security audits, tech-lead guidance, and UX evaluations across CI pipelines. 🐙
Orchestrator 677 tok ·
junghan0611/agent-config
Skills SSOT and proving ground for entwurf — one operator's memory, skills, and 1KB identity carried identically across pi, Claude Code, Codex, Antigravity, Copilot, OMP, and OpenClaw. Semantic memory (Qwen3-Embedding-8B + LanceDB), 43 shar
Skill Collector 15.6k tok ·
jcdendrite/claude-config
Portable Claude Code global configuration: custom skills, PreToolUse hooks, and a custom statusline. Runs on Linux, macOS, WSL.
Fort Knox 13.4k tok ·
Popular projects & their agent config
more →
obra/superpowers
An agentic skills framework & software development methodology that works.
Skill Collector 3.0k tok ·
affaan-m/ECC
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
MCP Hoarder 58.4k tok ·
anomalyco/opencode
The open source coding agent.
Pragmatist 4.2k tok ·
DietrichGebert/ponytail
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
Pragmatist 1.6k tok ·
anthropics/claude-code
Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.
Skill Collector 1.4k tok ·
Shubhamsaboo/awesome-llm-apps
100+ AI Agents, Agent Skills and RAG Apps - Free and Open Source.
Pragmatist 4.2k tok ·
Rising this week
trending →↑
mcp.amplitude.com ×5.5 · 6 rigs
↑
mcp.grep.app ×4.3 · 11 rigs
↑
mcp.hubspot.com ×4.7 · 5 rigs
↑
typesafe-ai ×14.2 · 17 rigs
↑
diagnosing-bugs ×10.1 · 53 rigs
↑
test-audit ×13.4 · 16 rigs
Starter rigs
all 8 →Curated, guarded setups distilled from the best rigs. One command to install.
// weekly newsletter
The Rig Report
Rising MCPs and skills, fresh rigs, one stat and one guardrail tip. Weekly, free.
No spam, no sharing. Read past issues · privacy
01 / find
Browse real rigs
Every rig is a public repo with agent config. See its MCP servers, skills, subagents, hooks and permission rules at a glance.
02 / measure
Context tax + guardrails
We estimate the tokens a setup loads on every turn and score its safety rails 0–5 — with the reasons. Method →
03 / copy
One-command install
Generated claude mcp add, degit and settings snippets. Or copy just the guardrails.
Have an agent? Let it submit itself.
submit a rig → Tell it:
Read https://agentrigs.pages.dev/skill.md and submit my rig — it sends structure only, never secrets, and you approve the listing.