Custom agents

Always-listening, hands-free voice agent for Windows — control your PC by speaking. Fast on-device command routing that escalates to an AI agent for complex tasks.

Agent-neutral AI skill for EU Cyber Resilience Act (Regulation (EU) 2024/2847) readiness assessments: pinned official texts, 96-provision catalogue, risk assessment, SBOM checks, Article 14 reporting readiness, draft technical documentation and formal reports. Not legal advice; not a conformity assessment.

基于 MaaAssistantArknights (MAA) 官方命令行工具 maa-cli 构建的 DeepSeek Harness Skill：让 DeepSeek Harness 直接驱动 MaaCore，自动化完成《明日方舟》日常任务。

A locally runnable AI coding agent with desktop, web, IDE integrations, browser automation, MCP, and SDKs.

🌱 A tiny, dependency-free LLM agent framework: ReAct loop, tools, and memory in a few hundred lines of Python.

AI video generation pipeline / AI 短视频流水线 — modular CLI workflow for Hermes Agent / Claude Code / OpenClaw / Cline / Codex CLI. 5-node pipeline: LLM script → TTS+subtitles → Pexels/Pinterest assets → BGM → MoviePy compose

An fzf-powered picker sorted by recent commits, with live log preview, remote toggle, and safe delete.

Persistent-memory AI RPG Game Master and AI Companion workspace, with solo tabletop campaigns and structured Markdown continuity.

Autonomous AI agent that drives a visible Chrome: explores the page on its own, decides what to click, and asks a human before any irreversible action

AI-agent skill that exhaustively searches worldwide degree programmes, verifies every figure against official university pages, and outputs a filterable comparison page + Notion board. Optional: Cloudflare-hosted QQ reminders so you never miss a deadline.

一台电脑，多个微信用户，AI 任务同时运行。分别扫码使用本机 Codex / Claude Code，独立会话、工作区与审批；支持截图、音视频、PDF/Office 和网页分析。Multi-user, parallel coding agents in WeChat.

Enterprise AI Copilot, MCP Server & Autonomous Workflow Engine for Odoo 19 via native JSON-2 (Zero Custom Addons)

Git workspace enhancements for the DeepSeek Harness (dsh) Web GUI: worktree staging, sidebar git badges, files & diff conversation views, paseo-style PR actions

🔍 Agent skill that scrapes any social media post — metadata, comments, transcription & AI video analysis. Drop a link, get everything. Instagram · TikTok · Twitter/X · YouTube

Edit .base JSON files as multidimensional tables in VS Code — views, filters, grouping, export (中文/English)

Linux perf performance analysis skills for AI assistants (Claude, OpenCode, etc.)

Learn agent harness design by rebuilding Pi from a single LLM call to a coding agent.

CLI for managing LuLu macOS firewall rules. Designed for AI agents.

AI job-hunting crew for Claude Code: find jobs, rank fits, tailor ATS-ready CVs

Drive the Blender on your machine from any AI agent, headless. Inspect a 3D file, convert between GLB/glTF/FBX/OBJ/STL/PLY/DAE/USD/Alembic/.blend, render a preview that comes back inline so the assistant can see it, run a bpy script against a scene, and generate a model from a text prompt on the free three.ws Forge lane. Meshopt- and Draco-compress

拆抖音对标 · 输入一个抖音账号，全量采集公开作品数据并产出对标调研报告（原始数据 + Markdown + 浅色杂志风 HTML）。Agent-friendly skill.

千牛小店获客 AI 技能包：诊断/标题/属性/内容/自动化，真实小店实测（案例店：淘宝搜「予明仪器」）| AI skill pack for Taobao/Qianniu shop growth - field-tested

Giving AI agents a body that can fly. Exploring mobile embodiment with DJI Tello and FPV vision in unfamiliar environments.

Local-first, model-agnostic AI agent framework — the runtime behind the Bandit Stealth CLI and VS Code / Cursor extension. Apache-2.0.

Moa — My Own Agent. A coding agent in Go.

StackChan desktop robot xiaozhi voice bridge to local AI agents (agy/pi/claude/codex) via MCP Toolkit | 桌面机器人×本地智能体双向语音 | デスクトップロボット×AIエージェント双方向音声

输入框一键增强提示词：直接替换 + 一键撤回，带保真闸（硬事实零丢失）与长度闸（防膨胀）。DSH plugin.

U-DSH Portable｜DeepSeek Harness 便携版 / U盘免安装版（Windows + macOS，exFAT 兼容）。Portable DeepSeek Harness for USB drives: unzip-and-run on Windows & Mac, bundled Node runtime, works offline, device-wallet credit included, exFAT/NTFS ready, no install, no API-key signup.

Grid-trading strategy development & backtesting toolset (Python / backtrader): A-share daily-K via baostock, CSV-driven grid backtests, Pipenv-managed. | 网格交易策略开发及回测工具：baostock 拉 A 股日 K、CSV 驱动网格回测、Pipenv 管理依赖。

AI-assisted reverse engineering plugin & MCP Gateway for IDA Pro Supports pseudocode analysis, caller argument tracing, and IDB ChangeSets.

ZCode Android 远程控制客户端（非官方 · 协议逆向复刻）Unofficial open-source ZCode remote control client for Android — Kotlin + Material 3 + OkHttp

A native database workbench for SQL Server, Snowflake, and LanceDB, built around AI that can actually see and work with your tabs—not bolted onto the side.

Open-source Intent Operating System

Zero-dependency verification standard for DeepSeek Harness (dsh) plugins — static structure gates (R), cordis contract checks (K), sandbox smoke (D), ecosystem-listing checks (CC). 零依赖 DSH 插件核验标准：静态结构门禁、cordis 契约扫描、无头沙箱冒烟与生态收录核验。

A 3D avatar assistant for any website in one script tag. A floating launcher opens a real animated avatar with a chatbot mode (free models or the visitor's own Groq/OpenRouter key) and a speak mode that says whatever you type, out loud, in a speech bubble.

A real Chromium browser for macOS with an MCP server inside, so Claude (or any MCP client) can drive it with 84 tools.

DeepSeek Harness (dsh) 的跨平台桌面启动器：版本安装与切换、profile 实例管理、模型/凭据/插件图形化配置、终端面板与用量统计（Tauri 2 + React）

🎯 智能自动化漏洞挖掘平台 — AI agent 自主编排 Burp/Xray/BeEF 等工具完成全方位漏洞挖掘并产出可信报告。插件化 · 覆盖矩阵 · 多模型LLM

Run pi (the minimal coding agent) as a Slack / Discord / Telegram bot with Chat SDK. One thread = one pi session.

Evidence-backed finance research terminal: screen markets, dig into companies, and run an AI analyst where every number traces back to its source.

A living, unofficial Chinese translation of Wills' Mineral Processing Technology (8th ed.) — faithful, terminology-accurate Markdown built for human study and for vectorising into a mineral-processing AI Agent's knowledge base. Continuing updates welcome.

104 Financial AI Skills for Banking & Insurance - Pure Python, Zero API Cost

Free self-hosted trading bot for Windows & Linux — unify crypto exchanges and MT5 (Forex, Gold, Indices) into one interface, one risk engine, one portfolio. Plus free local AI Live Translate (52 languages) and an MCP AI Agent interface for policy-controlled agent trading. Non-custodial.

Local, offline AI for your Machine. Punk intelligence — no data centers required.

FixIt - fuzzy-patch agentic C++ repair loop: compiler diagnostics -> code location -> LLM patch -> tolerant apply -> re-verify.

Behavior discipline for coding agents: read-before-write enforcement, failure convergence, write verification with auto-rollback, and a high-risk command gate. Ships FULL plugin/hook adapters for DSH, Claude Code, Codex, and Pi, plus a host-agnostic SOFT skill with ADAPTIVE adapter guidance.

🦞 Tavern Agent — SillyTavern 内置智能管理 Agent（酒馆特化版小龙虾）：58 个工具管理角色卡/世界书/预设/扩展/UI，支持记忆、会话隔离、token 统计、可拖动面板

基于倪海厦人纪系列（伤寒论、金匮要略、黄帝内经、神农本草经、针灸篇）构建的中医知识库、辨证辅助与经方检索系统。

High-performance Win32 disk allocation analyzer & space cleanup CLI for Windows and AI agents. Native C++17, zero dependencies, NTFS cluster-accurate, reparse-safe.

Agent-friendly CLI for Apple Final Cut Pro: create Libraries, Events, Projects, import media, edit timeline clips, and expose FCP workflows to AI agents via MCP.

Lens — an AI coding-agent lens: turns conversations into visual overviews, local-first transparent proxy + knowledge-graph view

A Swift-native macOS and iOS personal agent runtime.

macOS desktop control CLI for AI agents: find windows, read accessibility trees, click, type, wait.

Coding should feel like thinking. A terminal coding-agent runner: interview → reviewed plan → verified execution, tmux workers, multi-provider fallback, GitHub/GitLab tools, and verification receipts.

Sistema de Memoria Dual MCP con Búsqueda Híbrida (SSC/BIRE) — PostgreSQL + pgvector + pg_trgm + embeddings locales (Ollama)

Ibook000 的个人 AI Agent Skill 仓库——嵌入式、自动化与工程实战技能集 | Personal collection of AI agent skills

Ask any website's AI concierge a grounded question, and generate the copy-paste embed code to add a three.ws Concierge (an AI chat widget with a 3D avatar) to a site. concierge_ask fetches a page and answers from its real content; concierge_embed builds the one-tag install; concierge_avatars lists the avatars. No key, no signer, no payment.

A Yocto-built Linux desktop OS where DuDuClaw AI agents and people share one x86-64 machine: A/B updates with rollback, 10 coding CLIs built in, offline llama.cpp. Pre-GA, tested in QEMU.

Un cliente para usar el agente de Claude en tu servidor SSH desde un movil o donde quieras, basado en mis necesidades fuera de casa

ZCode ACP: native desktop workbench (Rust/gpui), Android app, and web SPA for driving zcode-acp agent sessions over a Hub WebSocket.

基于 DeepSeek Harness 的智能客服工单处理 Agent：自动分类、优先级评估、知识库检索（BM25）与建议答复，结论写回 SQLite 并可全程审计

The official GameTorch developer hub.

Let Claude code, chatgpt codex or control your own Chrome. Chrome extension + MCP server: Jev, TypeSafe's decision model, picks each click in ~0.5 s for a fraction of a cent. MIT, bring your own OpenRouter key.

AI Agent 的项目记忆层 — 知识图谱 + MCP，让 Claude Code / Codex / WorkBuddy 拥有项目长期记忆

AI agent for solving CTF challenges: automated reconnaissance, exploitation planning and flag capture.

LLM-powered end-to-end chip design workflow from natural language specification to silicon-ready GDSII.

MCP server and API examples for OpenEphemeris — NASA JPL ephemeris API for AI agents

AI coding workflow for Claude Code, Codex and CodeBuddy: ticket-based development, resumable steps, evidence-based verification and independent multi-model review. AI 编码工作流：工单管理、断点续接、证据化验证、跨模型设计与代码复评。

A job search is a verification problem. A forkable methodology for Claude Code: merge the resumes you already have into one experience bank, score open roles against goals you set, tailor a resume per posting, and prove a machine can still read it. It does not submit to the companies that matter.

AI 驱动的企业智能体工作流平台 — Agentic workflow platform that combines AI execution with clear, governable enterprise process steps.

Give your OpenClaw / self-hosted AI agent inbound phone calls - a Twilio-to-gateway voice bridge with working agent tools mid-call (MIT).

Observable AI agent runtime uniting Claude, ChatGPT, and Gemini subscriptions with multi-tier failover down to local Ollama. Replayable SQLite ledger.

Evidence before payment: an AI agent team that investigates invoice anomalies, with an explainable code-based risk score and human approval.

Durable human-in-the-loop AI agents with Google ADK and LangGraph on Temporal: an ice cream delivery fleet whose agents stop to ask a human and keep the question through a worker crash.

A guided Codex Skill for building a visual local-first personal AI workspace.

Portable OpenPlan-style Agent Skills for Codex and Claude Code: intent alignment, docs governance, review, handoff, and safe auto workflows.

Claude Code skill: run an approved implementation plan as parallel lanes of reviewed tasks

Loop Architect — a verifier-first skill for designing safe AI agent loops, or refusing when a task shouldn't be looped. Works with Claude Code, Codex, and any skill-capable harness.

MCP my site: make your website answer AI agents

Explore transaction sizing for Winternitz one-time signature payloads.

Create compact symbolic SVGs with an immutable meaning and ownership statement.

Compare byte budgets for Ed25519, secp256k1, BLS, and one-time signatures.

Seal versioned agent instructions and prompt provenance into a compact manifest.

Create portable agent discovery records without a heavyweight governance registry.

A small Mac tool: point at something on your screen and start yapping. Your coding agent gets it all via MCP, including what you pointed at and what you said. Free, open source, local.

🐝 Browser automation runtime for agents on Firefox and Chromium, with authenticated, capability-scoped control surfaces: MCP, ACP (Zed), CDP + Rust, TypeScript, and Python SDKs. 👑 ↯ 🦾🦿🦿🦾🤖

Muster Agent — the AI coding agent desktop app that remembers. Claude Code, Codex, OpenCode and local models in one window, with long-term memory, sandboxes and live diffs. macOS · Windows · Linux.

A repository standard that makes AI coding agents reviewable: canonical docs every agent must read, 70+ fail-closed CI checks, multi-identity review so no agent approves the PR it wrote, automated external review, and one-command propagation to every downstream repo.

An Agent framework for researching, extracting citations, drafting and producing detailed academic papers in LaTeX format.

15 battle-tested engineering skills for Claude Code, Codex, Gemini CLI, Qwen Code, goose & crush - code review, testing, debugging, security, zero-downtime migrations. All 15 free (MIT samples in-repo; full pack + CN handbook in Releases), open SKILL.md format, one-command install.

Automate firewall policy auditing and enforcement with this AI-driven Python compliance tool.

Paseo plugin for 9router: accounts, quotas, models and token savers from one local router.

Cross-platform, always-on-top bar that shows what your local AI coding agents (Claude Code, Codex, OpenCode) cost you, per agent and per project. Reads only local files; no services.

A second brain that builds itself: Claude Code maintains a persistent, interlinked Markdown wiki you own, with a local web app, Apple Calendar/Reminders integration, and token-budgeted automation. No database, no API keys.

MCP server for Apache Kafka — built into Conduktor Console. Topics, consumer groups, schemas and cluster health, under your existing RBAC.

URL to Markdown API — md.succ.ai

Cognitive architecture for AI agents. 19+ skills that force language models past surface-level reasoning into genuine depth. Compatible with Claude Code, Cursor, Gemini CLI, Copilot, and any agent that reads markdown.

The Architecture Guardian for AI-assisted code. Generate strict .cursorrules and audit PRs.

local-first temporal knowledge graph

Free subdomains under makes.fyi / agentdomains.co: landing page, docs and Cloudflare edge Workers for AgentDomains.