Trending Repos

A tool to improve the efficiency of coding agents by reducing the number of tokens used.

Why: It can lower the cost of using AI coding agents by about 31% without affecting their correctness.

Image generation (gpt-image-2) and GPT-5 subagents for Claude Code — through the Codex CLI login you already have. No OpenAI API key.

Open-source Agent Skill that drives the BaoCut macOS app CLI (transcribe · subtitle · translate · cut) from Claude Code, Codex, and other agents

Universal, model-agnostic operating harness for AI agents (Claude, Codex, Gemini, …) — a lean core + work-type profiles assembled by one setup script.

Codex skill for auditing and tuning Codex Desktop context/tool surfaces

Self-hosted, censorship-resistant proxy server with an all-in-one admin panel for Xray-core, sing-box, Hysteria2, and AmneziaWG.

Spec-driven autonomous engineering workflow for AI coding agents: recon → plan → build → validate → ship — with TDD, adversarial validation, and anti-slop quality gates. Packaged as a Kiro Power.

Claude Code plugin: preserve and restore working state around /compact

A long-horizon benchmark harness: give a coding agent a real game engine, professional conditions and time, and ask it to build an open-world game. Harness only, no results.

Experimental loader for Ableton Live 12+ on Linux using WINE

My Pi coding agent setup: 17 plugins + 18 global skills + 2 MCP servers, with a one-line installer and tutorial README.

感谢马斯克 让我grok自由，免费注册grok 无数个号。然后cpa自由使用

atp — fountain-coded file transfer (RaptorQ over UDP/QUIC) that outruns tuned rsync on real networks. Verified end-to-end, fail-closed.

Git worktrees that set themselves up. One shell function.

A memory layer for Claude Code that learns as you work. Persistent, per-context memory as plain markdown in your own git repo. No server, no telemetry.

The All-in-One Orchestrator for Rathole & Nginx

A transferable, self-enforcing software-engineering methodology for AI coding agents — playbook, skills, contracted subagents, lifecycle hooks, and evals.

Make Claude talk like a human: 17 installable output styles for Claude Code, built on credited methodologies

A skill that helps AI agents talk to humans.

one-shot, AI-first macOS setup for Apple Silicon

Turn a fresh LXC into an AI homelab command center. One command installs Claude Code + a web UI, preloaded with skills, and a credential broker so Claude operates your TrueNAS/Proxmox/Linux over SSH without ever seeing a secret.

Tells parallel AI coding agents when their in-flight work would break each other — before either lands. Catches the case git merges cleanly.

Graph engineering made concrete: run long coding tasks as a graph of agent nodes (executor + clean-context supervisor, more roles planned), not one drifting loop. 从 loop engineering 到 graph engineering。Claude Code skill · MIT · EN/中文

🌶️ A product-thinking layer for vibe coders — grills a raw idea into build-ready docs through 13 expert personas. Claude Code plugin.

給 Claude Code、Codex 等 agent 用的 macOS 磁碟清理 skill：診斷「系統資料」佔用、唯讀掃描腳本、以及 10 個會造成誤判或資料損失的陷阱

Claude Code skills I actually use, open-sourced one at a time

A disciplined, delegation-driven operating system for AI coding agents. npx contractor-kit.

Long-horizon agent skill for Claude Code / Cursor / Codex / Grok — multi-task ledger loop (related or not), host-portable (re-send the prompt to continue), clean-context supervisor, verified gates. Markdown library (loop-graph · quest), not a framework.

Give Claude Code eyes 👁️ — a camera skill so it can SEE real hardware, displays and wiring: verify a rendered panel, check wiring before power-on, and catch bugs that live on the glass, not the logs.

A curated collection of 51 production-ready subagents for Gemini CLI. Drop them into .gemini/agents/ and let Gemini delegate the right specialist.

🌱 Skill éco-responsable pour Claude Code : sobriété numérique, audit d'éco-conception (RGESN, GR491, Green Software Foundation) et pratiques de sobriété IA

AgentVitals Checkup (/checkup) — an AI agent skill that gives your agent a professional health checkup: dual-axis Stability + Welfare scoring, a personality-style title, and a public cross-platform leaderboard. One-line install for Claude Code / OpenClaw / Codex / Coze. Bilingual EN/ZH.

Run Hermes Agent free 24/7 on GitHub Codespaces + Tailscale VPN. Step-by-step guide in English 🇬🇧 and Vietnamese 🇻🇳

Claude Code skill: reverse-engineer any motion reference into production animation code via frame-level dissection

Stop opening /usage — a Claude Code status line that does the weekly-limit math: today's share, sustainable burn rate, sleep-aware pacing. Pure bash + jq.

Use Claude Code's UI with OpenAI GPT-5.6 (Sol/Terra/Luna) as the engine — powered by your ChatGPT Plus/Pro subscription via a local Codex proxy. One-line install; plain claude stays untouched.

Drive Google's Antigravity CLI (agy) from Claude Code: system-prompt sharing + a research-method prompt prefix. CLI counterpart to the IDE bridge.

OpenConfig — best OpenCode + oh-my-openagent (OmO) + OpenRouter AI coding agent config. Sisyphus · Claude · GPT · DeepSeek · Gemini · MCP · Context7 · Exa. oc install → doctor → launch.

Drop-in skills that stop AI coding agents from shipping garbage — kill-slop, security-review, PR hygiene

Open-source Codex plugin for Sol-governed multi-agent coding, bounded AI code review, model routing, and LLM cost optimization

Claude Code skill that delegates implementation to GPT-5.6 Sol via Codex CLI, Claude plans and reviews the diff, Sol writes the code. Multi-model AI coding where the model that wrote the diff never grades it.

Build and ship a Codex Site with Convex as its realtime database and backend—using one reusable Codex skill.

🌀 The first adversarial-collaboration multi-agent Skill for Claude Code. 7 agents + 11 phases + Hook enforcement + 9 signal groups + dual-layer adversarial review. Inspired by NTCoding, agent-review-panel, AgentTeams. v1.6 — main+sub dual-layer adversarial review.

Beginner-friendly guide and scripts for deploying projects with Git, focused on practical Shell-based workflows, basic automation, and simple deployment best practices.

A pseudo-Fable mode for Claude Opus — Fable-style planning, verification, and multi-agent discipline as a Claude Code skill set. Unofficial; not affiliated with Anthropic.

Claude Code skill: build a GitHub profile README with self-hosted stats cards (Lambda + CloudFront)

Run GPT-5.6 Sol and Codex models inside Claude Code — private, localhost-only, one-command setup for macOS and Linux.

Production CI/CD pipeline template with quality gates, container security, SLSA provenance, cosign signing, and automated release management.

Run Claude Code with GPT models through Raine's claude-code-proxy.

Run Claude Science on Windows. A WSL2-backed installer and taskbar launcher for Anthropic's research app.

Protection against npm supply-chain attacks. One command: Socket Firewall on every package manager, minimum release age, ignore-scripts.

Minimal CLI tool to manage a Cloudflare WARP (wgcf) tunnel with an nftables kill switch that blocks all traffic unless the VPN is active, preventing IP leaks.

Production-grade recipe for DeepSeek-V4-Flash-0731 (284B MoE) on 2x NVIDIA DGX Spark: self-healing 2-node vLLM cluster, reboot-verified, tuned DSpark speculative decoding (~75 tok/s), full benchmarks, OpenAI Codex CLI integration

Interactive disk cleanup skill for Claude Code on Arch Linux. Scans package caches, dev junk, Docker leftovers and oversized files, then cleans only what you approve.

AI-powered Databricks platform engineering — provision workspaces, configure Unity Catalog, manage identity & governance, and set up networking across Azure, AWS, and GCP, all through natural language with Claude.

A cli tool to stream music in terminal

Behavioral guidelines for LLM tools (AGENTS.md). Installer wires Claude Code, Codex, OpenCode, and Kimi Code; Cursor/Cowork via paste. Principles 1–4 Karpathy-derived; 5 by maintainer.

One CLI for your whole machine: code worktrees, notes/brain, offline local AI, and Mac messages/calls/calendar — for you and your AI agents.

Build and ship a Codex Site with Convex as its realtime database and backend—using one reusable Codex skill.

Build a MetaTrader 5 trading bot and deploy it on Senti — starting from a template that already gets the hard parts right.

把 Mirasim 内置的 Anthropic 反代接给独立 claude CLI 和 Zed ACP agent — 动态端口发现，不复制凭证

🪄 Wave wand over repo. Curses removed. Opus behaves. Twelve spells that de-hex your Claude instructions for the Opus 5 era.

Caps Lock LED as a live status indicator for Claude Code — blinks while working, solid when done. Zero dependencies, no Caps Lock side effects.

A reusable Claude Code configuration for agentic coding and software engineering workflows. It provides global engineering guidelines, debugging and bug-fixing workflows, code review, security auditing, testing skills, and reusable developer commands.

🛡️ AI-powered reconnaissance framework unifying 130+ security tools — subdomain enum, vuln scanning, cloud/IoT recon, dark web monitoring & blockchain-audited reporting. GAN-driven subdomain prediction, distributed execution post-quantum encryption. Built for bug bounty, red team & DevSecOps.

Tencent Hy3 295B MoE (NVFP4) on 2x NVIDIA DGX Spark — TP2 over 200GbE, 256K context, 26 tok/s end-to-end. Tuned, benchmarked, agent-ready.

A personal engineering methodology: routed requests, wave-based TDD plans, evidence gates, subagent execution, durable memory.

GLM-5.2 QuantTrio (Unpruned ) on 4xDGX Spark (GB10): 327K-655K context or up to 5 concurrent agents, one receipe. TP4 + DCP + MTP

Master prompts to make Claude (and other AI models) review code for security, performance, accessibility, usability and quality — language-agnostic, with add-ons for 18+ stacks. Bilingual EN/ES.

A configurable status line for Claude Code - model, effort, workspace, and live context/rate-limit, claude usage monitor bars.

Discipline + security starter kit that keeps Claude Code (or any LLM agent) on course across sessions: repo-based memory that survives compaction, enforced hooks & permissions, and a phased workflow — consistent, traceable, safe. Clone it or install as a plugin.

Claude Code skill for searching royalty-free stock photos via the Pexels API

Agent skills that facilitate having agents do code reviews on open PRs and the main coding agent address all the gathered comments

给 Kimi Code 的 web portal（kimi web）加一个 Files 文件面板：右侧文件树 + 面板内预览。一个干净的 git patch，不改二进制，重启即还原。Adds a Files panel to Kimi Code's web portal.

Tell your AI coding agent: plz do it, then prove it

Autonomous agentic coding environment — an OpenCode-driven 9-agent crew that ships hands-off. AGPL-3.0 · use at your own risk.

⚡ AI Content Workspace Template

Autonomous audit-fix-verify loop for Claude Code. Run against 17 open source projects in 8 languages, 16 converged, 3 fixes merged upstream. Public journal for every run.

A workflow operating system for AI-assisted engineering. Task state, decisions, and plans live on disk as files, not in chat history, so context survives across sessions, tools, and restarts.

Automated DNS Resolver Platform for ISP & Enterprise Networks powered by Unbound & dnsdist

Two AI coding assistants (Claude Code + Codex) review each other's risky plans before either writes code. Install once, works both directions.

Benchmark AI code reviewers on YOUR repo. Mines fix commits for answer keys, grades HIT/DEFER/MISS, and staffs a panel of models that fail in different places.

Workflow-agnostic guardrails for AI coding agents: spec sync + risk-triggered engineering discipline for Claude Code and Codex.