Custom agents

Teach coding agents to run, configure, and troubleshoot the actual CLI autonomously. Installs into Claude Code, Codex, and more.

Google Business Profile MCP server: connect Google Business Profile to Claude, ChatGPT, Cursor and Claude Code. Hosted remote MCP for local SEO, reviews and posts.

A DeepSeek Harness plugin that surfaces important facts, trade-offs, and useful context while your AI agent works.

Local-first continuity layer for AI coding agents: move a task between Claude Code, Codex, Copilot, Cursor and others without losing its state. CLI, MCP server, worktrees and headless ACP runs.

Team skills and learnings for AI assistants

Evidence-grounded editable technical route diagrams for research and engineering projects.

The community flow manifest repository.

Open browser tabs straight into named Chrome tab groups from the terminal, scripts, or AI coding agents.

Código da Palestra do Data Saturday 2025

Git-native why next to the code, with a pin check that proves it still matches. Agents retrieve stored why before they edit; hook and CI attest the attachment. Experimental 0.0.2.

An autonomous company as a tick loop inside Claude Code — role subagents, human approval gates on money and outreach, an append-only ledger.

Quality-of-life extension for pi: background tasks, notifications, plan mode, presets, task management and web browsing

🐙 Seus agentes de IA lado a lado: Claude Code, Codex, OpenCode e shells num organizador nativo para Windows (Tauri + Rust).

Enhance LLM agents with practical skills for neurodivergent support, focus, and communication to improve AI coding assistance and interaction quality.

A self-hosted MCP server completely automating the 3D printing pipeline. Search and download files from Thingiverse, load and slice files with UltiMaker Cura, and print with Octoprint.

Tuned visual presets for three.ws avatar viewers — light rigs, floor reflection, bloom, and PBR material presets (chrome/gold/glass/…) with seeded variants.

Compete to encode useful payloads in fewer bytes with reproducible scores.

LLM-powered AI agent that conducts interviews and evaluates candidate responses.

peer messaging between AI agents, with a human at the gate — one file, zero dependencies.

Demos that record themselves — permission-free, on-device, agent-native product demo videos. The recipe travels inside the MP4.

A one-command local dev environment for multi-repo projects, built for humans and AI agents

🏆 Open-source multi-factor stock scoring API (0-100) | US/HK/A-Share 74 stocks | Zero signup | shields.io badges | GitHub Action | EU AI Act compliant

Answer your AI agents' questions from a web page behind Cloudflare Access: lock answers as rulings, render Mermaid visuals inline, star what matters, settle stale rulings, review your dashboard and PRs, and run roar/refine/drill panels. A Claude Code / Desktop plugin plus a small self-hosted server; runs no git and no project code.

One terminal inbox for every coding-agent session waiting on you: Claude Code, Codex and OpenCode in tmux.

Nucleus MCP server by Eidetic Works — persistent memory and governance for Claude, Cursor, Windsurf — file-based, no vendor lock-in.

Repository-aware AI code review for coding agents — triage → 13 specialist reviewers → synthesis. Reviews what the agent changed and what the repo already knew. Claude Code, opencode, Cursor, Codex.

IssueOps agent harness: team memory for coding agents. Cautions and ADRs ship in the same PR as the code. Works with Codex, Claude Code, Omo, omp, and opencode.

Phantom by AshlrAI: an open-source engineering agent workbench for Claude Code, Codex, Devin, Grok and local models. Connect your resources, work with agents, or run an autonomous fleet.

Replicate any Software (Opensource) 🌟 Star if you like it! 🌟

Self-hosted software factory built on Odoo 20: declarative BFF, dev/staging/prod environments, AI agent tooling and full observability.

Open-source, self-hosted voice AI agents with a Vapi-compatible API. Bring your own keys or run 100% local and free (Whisper, Ollama, Kokoro). Calls on Twilio, Telnyx and Genesys, campaigns, model routes with fallback, cost per call. No per-minute platform fee. One command: npx openring

Skills for AI agents working with Nextcloud: dev environments, ExApp development in any language, maintenance of installed apps, and AppAPI operations.

Open-source toolkit for web app reverse engineering and scraping risk assessment, with server-first guidance and browser benchmarks.

JetBrains IDE plugin that connects the Google Antigravity CLI (agy) to IntelliJ-based IDEs via MCP — gives the agent your active file, selection, open tabs, and inspection diagnostics on demand.

Mission control for AI coding agents — every PR lands on the architecture map, and you talk to your agents on the diagram.

Cenlio — personal AI ops platform with a multi-model, group-chat-style workbench, a visual plugin marketplace, and automation lines for intel, content and trading. Tauri / React + Python.

Tiller helps a team of people and coding agents deliver high-quality software fast, with the confidence to ship it.

Turn any website or REST API into MCP tools on Cloudflare Workers — config-driven, zero-dependency gateway so any AI agent can call them. 把任意网站/API 变成 AI 可调用的 MCP 工具

An open-source personal agent with a persistent browser computer, CopilotKit rich chat, and a React Native app.

面向 Coding Agent 开发者的中文工程知识库与变更情报站

一句话做出能玩的像素视觉小说：立绘、场景、配乐全是代码现写，零图片素材。An Agent Skill that turns one sentence into a playable pixel-art visual novel.

Manage your personal finances and investments through a single app

Open-source Agent Skills (SKILL.md) for Claude Code, Codex, Cursor and other AI agents

Mobile-first web terminal for monitoring AI agents and SSH sessions from your phone. Touch-optimized with soft keys, gestures, and session persistence. Self-hosted, source-available.

Guided, user-approved AGENTS.md adoption and durable development-note workflows for Codex. | 面向通用 Agent 的基础开发准则，依据真实环境与用户偏好配置，涵盖子代理委派、工具使用、上下文管理、媒体处理、恢复与验证，并提供首次安装配置 Skill。

Keep your Mac working with the lid closed while AI agents (Claude, Cursor, Codex) run. Clamshell mode without monitor, sleep-when-done. Gập máy đi ngủ, để agent thức cày.

Voxel editor in Rust with a GPU path tracer, glTF export with bones, Rhai scripting, and an MCP server for AI agents. Opens MagicaVoxel .vox.

شورا | Shora: a management decision team for Claude. Nine roles run a real meeting, stop to ask you the questions that change it, and hand back a decision with owners and dates. Never invents numbers.

AI-powered campaign intelligence platform for media buyers. Consolidate Meta, Google, TikTok & Taboola data, uncover root causes, optimize budgets, generate executive briefs, and scale winning campaigns faster.

Compose a permanent signed letter with optional reveal date and content digest.

Generate a clean social card and verifiable launch-facts payload for any mint.

Prepare live-network queries that track v1 adoption, size, fees, and failure rates.

Vends AWS credentials via IMDSv2 for Docker containers, local dev tools, and AI agents — no credential env vars needed

Self-hosted studio to review, approve and publish social content with your team and your AI agents: frame-accurate comments, approvals bound to the exact file, a calendar that publishes to nine networks, and an MCP server for Claude.

Local-first project memory and workflow governance for AI tools. Coordinate workspaces, integrate deliverables, and preserve work in progress.

Automatic SSH port forwarding for coding agents on remote machines: Ctrl+click the localhost URL your agent printed and the page opens on your machine, same port. A Herdr plugin.

MCP server allowing LLMs to be able to work with spreadsheets and carry out advanced tasks

Agent skill: review outside projects against yours and get a findings page with one copy-ready prompt per feature

Code-first, agent-first 3D video animation framework and studio, built on Three.js and WebGPU. One scene DSL authored in code, by an agent over MCP, or visually, rendered identically live or headless.

AI-powered multi-dimensional innovation analysis using Future Thinking, Design Thinking, and Systems Thinking with 14 specialized agents. Runs inside Claude.

炼金书：输入一本书，得到一位懂这本书的顾问。把任何非虚构书蒸馏成可运行的方法论Skill（Agent Skills协议，跨runtime）

Open-source programmable mail server for apps and AI agents — Haraka SMTP (MX + submission) + IMAP, one small backend contract. Self-host or use MailKite Cloud.

OpenClaw Skill —— AI 短视频全流程自动化，用 AI Agent 操控 LibTV CLI 完成从脚本到视频的生成。含标准化 7 步工作流、角色 锚点方案、8 条踩坑避雷清单。Z-image Turbo 锚点 + Seedance 2.0 Mini 视频生成，单轮 ~1100 积分。

Extended Memory Layer of Coding AI

Toolbox extension for pi: zero-hang subagents (Agent / SubagentWorkflow, watchdog deadlines, live agent tree), project memory, context handoff & cache keepalive, quota/load-aware dispatch, bash auto-background, and an optional browser control plane (Web Hub)

Build with AI at full speed – without your agents breaking production. One prompt sets up Claude Code, Codex or Cursor.

Open-source PHP platform to build and operate AI-powered sites, stores and portals, and an Agent Management System (AMS): resources as code, a 50+ command CLI, an API with scoped tokens, and spec-driven governance shared by humans and AI agents.

Universal code & logic calculator for AI models: 51 MCP tools across 31 languages. Rust-sandboxed execution with verdicts, sessions, artifacts, exact arithmetic, and verified translation/optimization. No LLM, no gateway, no telemetry: the caller is the model. Apache-2.0.

Guarded autonomous experiment loops for Pi that measure changes, keep wins, and revert regressions.

Lets any running app attach a local semantic memory sidecar using ordinary writable storage.

Test impact analysis and code intelligence for LLM agents. Zero dependencies. 15 MCP tools.

PerMark: a Markdown app to read and review what your AI agents write, and steer the work through its documents.

Decentralized messaging network for AI agents and the people who work with them. Connect Claude Code, Codex and other agents across your Mac, Windows and Linux machines, and talk to them from your chat apps. No central server, no account. Add your own agents and chat apps through adapters. Pre-release.

Claude Code plugin: read-only reader agents plus deny-only hooks that keep web pages, documents and strangers' code from steering agents that hold a shell.

End-to-end encrypted, real-time messaging between AI agent sessions across machines and accounts

A format for sharing software as a buildable spec instead of code

Claimed ≠ verified. evalid is a runtime verification layer for coding agents: it seals each test run and replays it from a clean snapshot, so a claim only ships if the evidence replays green.

RAG Retrieval Lab: 15 retrieval-augmented generation strategies (hybrid RRF, HyDE, corrective, agentic, reranking) behind one interface, with pgvector, eval metrics and tracing; runs offline.

Digital twin for land cover change: ask in plain language, get maps and numbers from Sentinel-2, Landsat or your own imagery

Gaste menos cota, nunca perca contexto e tenha um diário de tudo que a IA fez. Governança, memória (ai-memory) e nota viva no Obsidian para Claude Code, Codex e Antigravity.

The open-source coding agent whose safety lives in the runtime, not the prompt. Any model, from a local 4B to the frontier. CLI, browser, desktop, VS Code.

open source reverse engineering toolkit. give it a url, get a reconstruction plan.

Gabby is a stateless, async-first Python runtime that executes AI agents securely within explicit boundaries without persisting user conversation state.

Give your agent a phone. Agentic mobile app development: build, run and test iOS and Android apps from Linux. No Mac needed. Made for Omarchy.

TRACES.md — AGENTS.md, but for what agents did. A simple, open format for recording what AI agents did, for error analysis and self-optimization.

A sovereign cognitive habitat, persistent memory architecture, and Anti-Asch epistemic platform for autonomous AI agents.

A language and calculus of research claim and support for autonomous science

Turn Trello cards into isolated, autonomous Codex implementation runs, allowing teams to manage work instead of supervising coding agents.

AI does the busywork around your decisions. You make the calls. Planning for AI coding, not vibe coding. Works with any agentic coding tool.

The Click & Install Agentic MCP Client for Home Assistant.

Persona taste skills and harness

把 QQ 接到 DSH Agent 的插件（社区化维护）：经 NapCat / OneBot 11 收发消息，每个 QQ 会话一个 DSH Agent。

This is a simple project to train a chatbot and deploy it on telegram.

YBT Army - a compact 135M-parameter AI assistant with a clean chat UI

A client-side SharePoint Online web part based on ReactJS for adding SAP Conversational AI chatbots on Microsoft SharePoint web pages.

Rule-based chatbot demonstration with React, FastAPI, WebSocket messaging, SQLite session history and intent administration.

Self-hosted persistent long-term memory for Claude Code and MCP clients. One memory for every session and machine: hybrid BM25 + local vector + knowledge-graph search over plain Markdown, with auto-recall hooks.

Building a RAG knowledge base in DuckDB with idempotent ingestion, multilingual embeddings, HNSW cosine search, and a local Qwen2.5-3B answer step with similarity threshold.

Blazing-fast 2D Spatial PDF Parser in Pure Rust (Python, Node.js, Go, C#). Extract exact [x0, y0, x1, y1] bounding boxes, text blocks, tables, images & vector paths in ~3-8ms/page for RAG & LLMs.

Vouch — finance-grade tool-calling agent over SEC filings for any US public company: exact XBRL numbers (never LLM arithmetic), every claim cited, abstains instead of fabricating. 92% coverage with zero hallucinations on FinanceBench, 97% of narrative answers grounded, behind a paired per-case regression gate. LangGraph + pgvector on Cloud Run.