How real projects configure their agents

24,785 repositories keep their own agent configuration in the open: a CLAUDE.md, a rules file, instructions for Copilot, a skill the project runs on itself. These are not mods to install. They are worth reading for what they reveal — what a maintainer thinks an agent gets wrong.

Building agents8,489Claude Code6,881Git and workflow3,605Data and AI3,580Backend and APIs3,556Languages3,486Planning3,323Code review2,893Testing2,681Writing and docs2,477DevOps and cloud2,477Memory and context2,375Research2,318Frontend2,308Security2,170Project management1,960

All Claude Code32,505 Codex24,785 OpenCode23,702 GitHub Copilot3,719 Gemini CLI2,034 Cursor1,771 Windsurf202 Kiro114

Project Stars Configures Files Always loaded
shuxueshuxue/SpexFlow Node-based workflow for orchestrating codebase context and specifications for AI agents. 163 Codex, OpenCode 1 341 tok
Khan/wonder-blocks khan.github.io React components for Wonder Blocks design system. 163 Claude Code, Codex, GitHub Copilot, OpenCode 11 14,901 tok
artem-sedykh/mini-humidifier artem-sedykh.github.io Minimalistic humidifier card for Home Assistant Lovelace UI 163 Claude Code, Codex, OpenCode 2 12,430 tok
vercel-labs/autoship CLI tool to automate changeset-based releases 163 Codex, OpenCode 1 225 tok
BV-Venky/excalidraw-architect-mcp Turn your architecture into a living, queryable knowledge graph - and render it as beautif 163 Claude Code, Codex 2 —
amplitude/builder-skills Open-source AI skills for product managers — battle-tested by PMs at Amplitude. 163 Codex, OpenCode 1 1,695 tok
calebl/ynab-mcp-server Model Context Protocol for YNAB (you need a budget) 163 Claude Code, Codex, OpenCode 3 684 tok
ppgranger/token-saver ppgranger.github.io Content-aware output compression for AI coding assistants. 36 specialized processors cut C 163 Claude Code, Codex, OpenCode 2 1,238 tok
konippi/servo-fetch A self-contained browser engine that fetches, renders, and extracts web content as Markdow 163 Codex, OpenCode 1 1,928 tok
NiceEval/NiceEval niceeval.com build eval for your agent in 10 mins 163 Claude Code, Codex, OpenCode 23 2,787 tok
nanvon/cc-bar Native macOS menu bar AI quota monitor and local usage dashboard. Tracks Codex, Claude Cod 163 Claude Code, Codex, OpenCode 3 3,348 tok
hermes-webui/hermes-android The best way to use Hermes from your Android 163 Codex, OpenCode 1 9,893 tok
diodeme/Gold-Band Desktop app for harness engineering, loop engineering, graph engineering—and whatever come 163 Claude Code, Codex, OpenCode 34 3,129 tok
educates/educates-training-platform educates.dev A platform for hosting interactive workshop environments in Kubernetes, or on top of a loc 163 Claude Code, Codex, OpenCode 9 5,977 tok
SAP/open-ux-tools Enable community collaboration to jointly promote and facilitate best in class tooling cap 162 Claude Code, Codex, OpenCode 6 7,823 tok
glowbom/glowbom-oss glowbom.com Build software like writing a book 162 Claude Code, Codex, OpenCode 2 303 tok
Dicklesworthstone/ultimate_mcp_server Comprehensive MCP server exposing dozens of capabilities to AI agents: multi-provider LLM 162 Codex, OpenCode 1 1,408 tok
hugodemenez/deltalytix deltalytix.app Trading journal with AI agents and dashboard with statistics 162 Codex, OpenCode 1 2,679 tok
alpacahq/alpaca-skills app.alpaca.markets Agent skills for Alpaca's Trading API and Broker API: drop-in SKILL.md files for AI coding 162 Codex, OpenCode 1 420 tok
RightNow-AI/AutoMegaKernel runinfra.ai An agent harness that compiles a model into one provably-correct, self-retargeting CUDA me 162 Claude Code, Codex, OpenCode 6 2,279 tok
sunyuzheng/lizheng-open-context superlinear.academy Source-grounded public context for 立正: Superlinear posts, selected comments, YouTube, Publ 162 Codex, OpenCode 1 916 tok
manavmishra/ZeroSlop zero-slop.ai Open-source Agent Skill that finds and removes AI slop from your writing 162 Claude Code, Codex, OpenCode 2 2,861 tok
Sstobo/Claude-Code-Game-Master Total conversion for Claude Code. Use RAG and the RPG ruleset apis to play a persistent ad 161 Claude Code, Codex, OpenCode 37 5,757 tok
TianGzlab/OmicsClaw tiangzlab.github.io Conversational & memory-enabled AI research partner for multi-omics analysis. CLI + Deskto 161 Claude Code, Codex, OpenCode 2 10,666 tok
21-DOT-DEV/swift-secp256k1 21.dev Swift cryptography library for Bitcoin and Nostr. ECDSA, Schnorr, ECDH, and zero-knowledge 161 Codex, OpenCode 1 3,188 tok
pandazki/pneuma-skills Co-creation infrastructure for humans and code agents — visual environment, skills, contin 161 Claude Code, Codex, OpenCode 15 3,488 tok
DataDog/pathfinding.cloud pathfinding.cloud An AWS IAM Privilege Escalation Path Library 161 Claude Code, Codex, OpenCode 7 8,249 tok
st0012/cctop cctop.app A keyboard-first menubar app to monitor and jump between AI coding sessions — minimum setu 161 Claude Code, Codex, OpenCode 10 2,147 tok
SerhiiKorniienko/bullshit-detector korniienko.dev Agent skills that fact-check the internet: claim-by-claim verification with sources and a 161 Claude Code, Codex, OpenCode 2 2,575 tok
farukkamcici/mectrics mectrics.app A lightweight, private system monitor that lives in your macOS menu bar. CPU, memory, batt 161 Claude Code, Codex, OpenCode 2 7,571 tok
ZhangShenao/harness9 zhangshenao.github.io Local-First · Lightweight · Feature-complete · Production-ready general-purpose Agent fram 161 Claude Code, Codex, OpenCode 13 19,184 tok
rldiao/mealie-mcp-server MCP server that exposes Mealie APIs to MCP clients such as Claude Desktop 161 Codex, OpenCode 1 557 tok
camilleroux/genart-skill Generative art skill for Claude Code — deterministic, hash-seeded, onchain-ready 161 Claude Code, Codex, OpenCode 1 763 tok
leonardsellem/codex-subagents-mcp Claude-style sub-agents (reviewer, debugger, security) for Codex CLI via a tiny MCP server 161 Codex, OpenCode 1 887 tok
workos/case 161 Claude Code, Codex, OpenCode 2 1,407 tok
vybenetwork/x402-client Client SDK for the Vybe x402 API. Pay-per-call USDC over HTTP and prepaid-credit WebSocket 160 Codex, OpenCode 1 1,910 tok
CortexLM/cortex Autonomous research network on Bittensor subnet 100, turning reproducible AI research into 160 Codex, OpenCode 1 2,574 tok
kryptobaseddev/cleo cleo-dev.com Agent-first task orchestration. Persistent memory. Multi-provider coordination. 160 Claude Code, Codex, Gemini CLI, OpenCode 6 10,498 tok
yeet-src/sigwire yeet.cx tail -f for signals. Every signal any process on the box raises — who sent it, who it hit, 160 Claude Code, Codex, OpenCode 2 7,313 tok
themanyone/voice_typing State-of-the-art offline (or networked) voice typing everywhere + text terminals (Linux or 160 Codex, OpenCode 1 1,007 tok
ifiokjr/monopi ifiokjr.github.io One-click setup for pi-coding-agent — extensions, themes, prompts, skills, and ant-colony 160 Codex, OpenCode 1 2,253 tok
aaif-goose/goosetown Steampunk geese run a parallel processing commune. Surprisingly effective. 160 Claude Code, Codex, OpenCode 13 1,254 tok
Makio64/threejs-cinematic-world-zoom cinematic-zoom.vercel.app A cinematic zoom on places in the world 160 Claude Code, Codex, OpenCode 2 2,609 tok
nexus-research-lab/nexus nexusos.cn Multi-agent collaboration platform for persistent, proactive AI agents across rooms, works 160 Claude Code, Codex, OpenCode 3 2,922 tok
chyinan/Kokoro-Engine Cross-platform virtual character immersive interaction engine 160 Claude Code, Codex, OpenCode 2 5,956 tok
first-tree-ai/first-tree first-tree.ai First-tree routes work to the right agent, gives it the same context your team has, and lo 160 Claude Code, Codex, OpenCode 2 2,114 tok
maplibre/maplibre-agent-skills Community-maintained agent skills for MapLibre GL JS — helping AI coding assistants write 160 Codex, OpenCode 1 1,218 tok
UR-SillyB/AirFerry Fully offline optical file transfer: browser-extension sender <-> Android receiver via QR 160 Codex, OpenCode 1 38,035 tok
DogInfantry/claude-skill-management-consultant-B1 Claude skill & plugin: turn Claude (or any LLM) into an MBB-grade management consultant. C 160 Claude Code, Codex, GitHub Copilot, Gemini CLI, OpenCode 4 665 tok
Strategic-Automation/violin strategic-automation.github.io Supervised, Hermes-native penetration testing profile with 35 routed playbooks, guarded ta 160 Codex, OpenCode 1 2,590 tok