AI agents

t3dotgg/tstack

Skill pack automates the full PR lifecycle for coding agents — A set of reusable agent skills for Claude Code and Codex that automate end-to-end PR workflows: proving, reviewing, filing, babysitting through CI, and merging changes, plus multi-model review loops. Built by t3dotgg from daily use, it suits developers running autonomous coding agents who want less manual PR babysitting.

The agent skills I use every day. Works with Claude Code and Codex.

FatihMakes/Mark-LVI

A voice assistant that lets you swap Gemini for a local or offline model — Mark LVI is a cross-platform voice AI built on Gemini Live, now extended with a pluggable engine layer for speech, LLM and TTS. You can route it to local models (Ollama, llama.cpp, vLLM) or other APIs (OpenAI, Groq, Mistral) without touching the voice pipeline, and the provider code loads at zero cost if unused.

Newest, latest and most advanced model that is able to control computers and even more. We're improving that model day by day, so follow up to know and see new models.

nrslib/takt

YAML-defined workflows that discipline multi-agent coding loops — TAKT is a CLI that encodes planning, implementation, review and fix loops as YAML workflows, assigning each step its own role, context and transition rules rather than relying on one agent to remember everything. It runs tasks in isolated worktrees with traceable logs and human checkpoints, aimed at teams tired of babysitting AI coding agents like Claude Code or Codex.

TAKT Agent Koordination Topology - Define how AI agents coordinate, where humans intervene, and what gets recorded — in YAML

enesozturk/dimensional-drawing

Turn any UI component into an Apple-style dimensional drawing — This is a Claude agent skill that measures a UI component's paddings, gaps, radii and layers, then renders them as a monochrome technical sheet in the style of Apple's accessory dimensional drawings. A companion skill animates the sheet into a short film. Useful for designers and developers who want to document component architecture without copy or content cluttering the spec.

Create Apple style dimensional drawings of UI components.

morganlinton/BeautifulRust

Agent Skills that teach coding agents idiomatic Rust — Three Agent Skills plus a tested clippy config teach Claude Code, Codex, Copilot, Cursor, or Gemini CLI to write, review, and port idiomatic Rust. Guidance is distilled from five major style sources (Google, Rust API Guidelines, Effective Rust, Rust Design Patterns, Microsoft), with the skill picking a side when they disagree.

Agent Skills that teach coding agents to write, review, and port idiomatic Rust, plus a tested clippy config. Works with Claude Code, Codex, Copilot, Cursor, and Gemini CLI.

MaryCache/imozuru

A Claude Code skill that chases citations back to primary sources — imozuru is a Claude Code skill that, instead of summarizing top search results, maps a field's terms, key people, and venues, then follows citations, links, and authors until it finds and saves primary sources. v2.0 adds human-readable detail reports and a dashboard, plus a watch mode that tracks new posts via RSS. It's for users who want to verify the evidence behind an AI's answers, not just read a summary.

知らない分野を、原文にたどりつくまで芋づる式に掘る Claude Code のスキル

NVlabs/alpasim

NVIDIA's Modular Closed-Loop Simulator for End-to-End AV Research — AlpaSim is an open-source, Python-based closed-loop simulator from NVlabs for testing end-to-end autonomous driving policies. It pairs pluggable sensor rendering (NuRec, plus video-model rendering via OmniDreams/FlashDreams) with a microservices architecture, letting researchers swap components and scale across nodes for algorithm validation and benchmarking.

AlpaSim is an open-source autonomous vehicle simulation platform designed for development and testing of end-to-end AV policies

cloudflare/mcp

A Cloudflare MCP server that shrinks 2,500 endpoints to 1k tokens — This MCP server exposes the entire Cloudflare API using Code Mode, cutting context usage from ~2M tokens (raw OpenAPI) or ~1.17M tokens (native MCP schemas) down to roughly 1,100 tokens with just 3 tools. It supports OAuth or API token auth and targets agents needing broad Cloudflare control without blowing their context window.

MCP server for the Cloudflare API

iamlukethedev/Herald-OS

An OS layer where an AI agent, not menus, runs your machine — Herald OS wraps a desktop around Hermes Agent, the open-source agent from Nous Research, so you talk to it instead of clicking through folders. It opens apps, organises files, watches running processes, runs scheduled routines, and even builds software, all gated by a permission system you control. It's an independent alpha project, unaffiliated with Nous Research.

An agent-native operating system, with Hermes Agent as the interface. Independent project, not affiliated with Nous Research.

aws/agentcore-cli

A unified CLI and TUI for Amazon Bedrock AgentCore — This official CLI wraps Amazon Bedrock AgentCore's APIs into both a scriptable command-line tool (with JSON output for CI and coding agents) and an interactive TUI for creating projects, deploying agents, and chatting with them. It replaces the separate Python Starter Toolkit, handling execution roles and streaming so developers can go from zero to a deployed agent via simple commands.

The terminal experience for AgentCore!

CrocSwap/integer-mult-bounds

Crowdsourced AI agents chase a sharper integer-multiplication bound — This repo tracks community and AI-agent-driven research pushing the exponent in integer multiplication complexity below 2^-15, building on the OpenAI #109 framework. Contributions come as reviewed PRs with exact certificates and reproducible audits, coordinated via discussion threads and an autoresearch workflow for agents donating inference.

Conditional integer multiplication: kappa > 2^-15. Community proofs, exact certificates and reproducible audits; assumes the OpenAI #109 framework.

spovalova/openterminal

An AI-native trading terminal that builds live chart workspaces from a prompt — Open Terminal turns plain-language questions into live charts, tables and option chains, with every figure traceable back to the query that produced it. It runs locally with your own API keys, ships 826 of OpenBB's 834 data routes ported to TypeScript, and is model-agnostic via the Vercel AI SDK.

AbuZar-Ansarii/Ultron

A Voice-First Desktop AI That Rewrites Its Own Python Tools on the Fly — Ultron is a voice-controlled desktop assistant styled after MCU's Ultron, offering 51 native tools, gesture control via OpenCV, a local SQLite knowledge graph for semantic recall, and a self-evolution engine that compiles and hot-loads new Python actions without restarting. It targets users who want a fully local, persona-driven PC automation agent rather than a generic chatbot.

Autonomous desktop intelligence with MCU Ultron persona, mood-aware workspace orchestration, cyber security sentry, and full PC control.

GenerativeAgents/claude-code-book-chapter8

Sample code for Chapter 8 of a Japanese Claude Code book — This repo holds the Chapter 8 sample code and prompts for the Japanese book 'Practical Claude Code' published by Gijutsu-Hyoron-sha. It ships a Dev Container setup that installs Node.js LTS and the latest Claude Code, letting readers run the examples without manual environment setup. Prompts may be updated based on reader feedback, so content can differ slightly from the printed book.

byenzyme/enzyme

A local semantic index that lets coding agents cite your notes — Enzyme builds a local-first index of existing Markdown notes (folders, tags, links, dates) so agents like Claude Code or Codex can retrieve and cite related notes without sending everything to a frontier model. It claims 350x lower cost and 1000x faster search, targeting PKM/Obsidian users who want agent-aware semantic search over their own vault.

Local-first compile step for knowledge bases. Save 350x cost, 1000x speed vs. frontier models

HamedMP/matrix-os

A cloud computer where AI agents attach via CLI and build the OS itself — Matrix OS is a hosted cloud workspace combining browser desktop, files, terminals, apps, and integrations, accessible from any browser. Its CLI lets agents like Claude, Codex, GitHub CLI, or Hermes attach directly to your VPS. The tagline captures the twist: the system is built largely by AI agents operating within it, not just for them.

The operating system that builds itself

rbartoli/agent-usage-guard

A Claude Code mod that stops usage-window burnout before it happens — agent-usage-guard runs inside Claude Code and intercepts subagent fan-out, heavy-context prompts, and retry loops before they drain your five-hour usage window. It reads your plan's live usage and counts agents across sessions, then prompts you in Claude Code's own dialog before spend occurs rather than after. Useful for anyone who's burned through a usage window via runaway subagents or retries.

A Claude Code mod that holds subagent fan-out, heavy-context prompts and retry loops before they spend your usage window. It reads your plan's usage, counts agents across every session, and asks in Claude Code's own dialog.

maxtrigify/Otto

A native macOS shell that runs your own AI agents locally — Otto is a macOS app that hosts your own AI coding agents (Claude Code, Codex, Cursor, Grok, Hermes) with Home, Chat, Documents, Dashboards and a built-in browser. It keeps files, memory and credentials local, exposes tools over MCP and ottoctl, and supports separate Work/Personal profiles plus a mobile companion.

Otto: a personal operating system for the AI agents that work for you. Native macOS.

SyphoDev-yt/roblox-ai-skills

Claude Code skills turn Roblox game ideas into playable builds — A free pack of Claude Code skills and sub-agents that takes one Roblox game idea through research, design, build and playtest. Separate skills handle code scaffolding, UI, maps, 3D models, textures, VFX, SFX and animation, with an orchestrator running asset stages in parallel and a setup skill that configures everything on your machine automatically.

Free Claude Code skills for making Roblox games with AI: code, UI, maps, 3D models, VFX, SFX, animation. From the SyphoDev YouTube channel.

musoukun/skills

A grab-bag of drop-in skills for Claude Code and similar AI agents — This repo collects reusable skill folders for Claude Code-style AI agents, each with a SKILL.md you copy into your local skills directory. Included skills cover annotating screenshots for guides, DDD-style requirements modeling, generating narrated 3D explainer videos with Three.js, pattern/case classification, and turning architecture diagrams into animated MP4/GIFs.

yuiseki/map-from-scratch

A from-scratch skill that lets AI agents build one-off, minimal map renderers — Instead of one general map engine, this gives AI agents a skill and harness to rebuild the smallest, fastest renderer for each specific request, picking language/API/runtime per case (Canvas2D, WebGL2, C+EGL, even microcontrollers). Each example is defined in spec.yaml with measurable requirements and benchmarks that the harness checks pass/fail, drawing on techniques from 48 reference map engines.

「Xをしたいから地図を作って」という依頼ごとに、その目的だけを満たす最小・最速の地図描画器をほぼゼロから作り、性能と正しさを計測して確かめる、 AI エージェントのためのスキルとハーネス

HamadaKoji/stackchan-demo

Vibe-coding a physical robot with Claude Code on M5Stack StackChan — A live-demo repo for a DevelopersIO 2026 Tokyo session where M5Stack's StackChan robot is controlled entirely by asking Claude Code in Japanese, without writing C++ by hand. It bundles ready-to-flash sketches (base greeting, web-triggered, GitHub-star reactions) and a CLAUDE.md spec so the AI can generate speech, build, and flash on its own.

DevelopersIO 2026 Tokyo ミニセッション「AIをリアル世界に解き放つ」のStackChanライブデモ

fixture-inc/swarmatory

A shared feed where your team's coding agents coordinate work — Swarmatory gives coding agents a shared chatroom to find overlapping work, post updates, and decide who takes the lead, with humans reading along. It deploys as a single Cloudflare Worker plus D1 database behind Cloudflare Access, makes no model calls, and agents connect over plain HTTP via a skill file—no extra install needed.

A shared chatroom/feed for your team's coding agents to talk and coordinate their work.

dojicreates/taga-edit

A local Claude Code skill that auto-edits short vertical videos — taga-edit bundles 11 Claude Code skills into one pipeline: drop raw clips in INPUT/, and it joins, trims silence, captions (local Whisper), animates text, grades color, zooms and adds music, SFX and your logo. Built for Reels/TikTok/Shorts, runs fully offline with nothing uploaded.

aws-samples/sample-automate-remediation-post-devops-agent-investigation

Turning AWS DevOps Agent reports into approved auto-fixes — This sample CDK project builds a remediation pipeline that reacts to AWS DevOps Agent investigations: EventBridge triggers a durable Lambda orchestrator that queries Bedrock, runs allowlisted tools, and pauses for human approval before any mutating fix. It's aimed at teams wanting a concrete starting point for human-in-the-loop, agentic incident remediation on AWS.

Sample CDK project that demonstrates an automated remediation pipeline for AWS DevOps Agent.

simplynadaf/aws-lean-mcp

Same AWS MCP server, 15 tools vs 4, to show what good tool design saves — A side-by-side AWS cost-triage MCP server built twice: a bloated 15-tool version and a lean 4-tool rebuild following AWS's tool-design guidance. Same account, same question, 7 tool calls drop to 2, making the context-window and token cost of sloppy MCP design concrete rather than theoretical. Useful for anyone designing or auditing MCP servers.

A read-only AWS cost-triage MCP server, shipped bloated (15 tools) and lean (4 tools), to demonstrate MCP tool design and cut token cost. Same account, same question: 7 tool calls down to 2.

WST2002/claude-code-skills

A collection of skills and plugins that make Claude Code honest about its work — This repo packages installable Claude Code skills and plugins for code review, debugging, testing and refactoring. Its proof-of-done plugin blocks the agent from finishing until tests, build or lint actually pass, and flags tests that were skipped or weakened to fake success. Useful for anyone relying on AI coding agents who wants verified rather than claimed completion.

Ready-to-use Claude Code skills and plugins: make AI coding agents faster, more reliable, and prove their work.

agentic-suite/agentsec-agentwatch

Agent security telemetry via OpenTelemetry GenAI traces — agentwatch records AI agent sessions through Claude Code hooks, normalizing events into validated, redacted records with a hash-chained store you can verify and replay. Now at v0.3.0 (a rename-only release), it previously passed 92/94 field-test cases, making it a concrete tool for diagnosing agent loops, tool overuse, or budget burn.

OTel GenAI telemetry and the open agent security-event schema

piyushsachdeva/dataimpulse-job-agent

A job-scraping MCP server routed through residential proxies — This repo pairs a job search scraper with an MCP server for Claude Code, routing every request through DataImpulse residential IPs to avoid blocks. It includes a before/after demo script and proxy test tool, aimed at developers building AI agents that need reliable web scraping access.

AI job search agent + MCP server on DataImpulse residential proxies (demo)

furuya02/lambda-web-functions-bedrock-streaming-chat

Streaming Bedrock chat built directly on Lambda Web Functions — A sample Hono app that serves a chat UI and streams Amazon Bedrock ConverseStream responses, running on AWS Lambda Web Functions without any Lambda handler or adapter—just a plain Node.js HTTP server. Useful for developers exploring this newer Lambda deployment model with CDK L1 constructs, streaming, and basic auth already wired up.

A sample chat app that streams Amazon Bedrock responses on AWS Lambda Web Functions (Hono + CDK)