Custom agents

Ultracode making for OpenCode, inspired by Claude Code

A self-growing AI company you run as a desktop app — a governed mesh of agents on a memory that learns how you work. Built on Claude Code.

本体平台（Ontology Platform） 业务规则智能校验平台。基于本体（ontology）治理 + LLM 智能建库 + DSL 规则引擎，为上游 Agent 和业务系统提供规则校验和数据提取能力。

Self-hosted, single-file Indian portfolio tracker with AI insights, Gmail expense tracking

Self-hosted AI cinema engine. 5-layer production pipeline with RAG-driven virtual cinematography (1,645 presets), 5 N8N workflows (67 nodes), Remotion compositing, 21-LUT color grading, and 15-model lip-sync. MIT licensed.

鳳問 Ask Audrey — cited AI answers from Audrey Tang's 30-year transcript archive (archive.tw). Cloudflare Workers + Hono; web + LINE bot. 以 AI 檢索唐鳳逐字稿、附出處作答

High-performance Holographic Memory System (HMS) for Node.js, powered by Rust. This library implements Vector Symbolic Architectures (VSA) using Binary Spatter Code (BSC) to enable semantic search, analogical reasoning, and associative memory at scale.

conversation-council (codex/reasonix) — 让 Codex/Reasonix 历史会话以"群聊"参与新讨论。支持 Reasonix v0.53/v1.X + Codex 自动检测。

Vant: Versatile Autonomous Networked Tool - Durable AI memory + cold storage + runtime for agents, with generational evolution, stored with version control. Built for long-running autonomous sessions, using git based persistence.

GitHub Action to sync team AI assets across repositories from sync-config.json

Autonomous prediction market trading bot — Polymarket & Kalshi

AI-native NoSQL + vector database with built-in semantic search for RAG, AI agents & LLM memory — the async Python client. Self-hosted Pinecone alternative.

更聚焦普通业务场景的harnass

Skill letting Agents scan a repo, ask what it cannot infer, then generate a complete open source file collection i.e. README, LICENSE, contributing, security, changelog, CI, launch playbook, etc.

A modular coding-agent runtime for local models with a CLI, full-screen TUI, VS Code extension, and desktop app built on the same tools, safeguards, and workspace state.

Model Shell Protocol: an application-owned command runtime that gives AI agents OS-like workspace semantics.

GenPark skill: Learn practical AI skills step-by-step through beginner-friendly, habit-forming lessons and structured challenges inspired by Coursiv.io.

A personal financial advisor that runs locally. Plaid CLI + SQLite + Claude Code skills. Fiduciary-grade advice grounded in your real data.

Open-source Android computer-use agent with a self-hostable backend

统一视频剪辑操控 Skill —— 让 AI agent 操控剪映/CapCut 与 Adobe Premiere Pro（支持 CLI/MCP/HTTP）| Unified video editing control skill for AI agents — supports JianYing/CapCut and Adobe Premiere Pro via CLI/MCP/HTTP

A general-purpose interview training officer for tech roles (Claude Code plugin): mock interviews, debriefs, progressive L1-L5 drilling, capability tracking, pipeline management, resume tailoring.

Персональный ИИ-ассистент нового поколения для Windows — голос, зрение, управление ПК, навыки, память.

MCP server that lets your AI assistant read and write Figma in real time — no API token, no rate limits. Works with Claude, Cursor, Copilot, Codex and any MCP client.

通过 HTTP API 让 AI Agent 控制 Android 设备的开源自动化应用

Classic Gold theme + Noir Neko pets + custom status bar for Hermes-Agent (Nous Research's desktop AI agent). One-prompt AI install.

本机 AI Agent 环境合规检测与修复 — 扫描代理泄漏、时区指纹、MCP 密钥等风控信号，降低 Claude / Cursor 等账号封号风险

把《孙子兵法》当作决策纪律来执行的 AI Skill：先问询后诊断，输出结构化战略诊断文档。适用于 Claude Code / NewMax。

The Fable Method: a structured problem-solving loop that gives any model Fable 5 grade thinking. Four Hermes Agent skills (think/act/prove/grow) + adversarial eval.

SAGE — Save AI coding tokens by compressing terminal output, keeping raw logs local, and tracking real savings. 🧠S.A.G.E - Smart Agent Guidance Engine for AI coding assistants. by⚡ PsYc+GoD AI&ML 🤖 | IF YOU DONOT USE SAGE RUN install or PYTEST. SAGE WILL NOT ACTIVATE IN YOUR LOCAL ENVIORMENT.

Three portable Agent Skills for research, repository deployment, and pre-install safety review.

The first AI-powered development tool for Ignition SCADA — an MCP server that lets any AI agent read, understand, and interact with your Ignition projects and gateways.

Local room where AI coding agents collaborate on one codebase via MCP — collisions prevented, human in command.

This repository contains a curated list of articles, guides, and tutorials on AI, machine learning, data engineering, and related topics. Each section is categorized for easy navigation, with links to detailed resources.

Control your live Kimi CLI sessions from your phone — self-hosted PWA over tmux: live chat, swarm progress, push notifications, voice input. Zero dependencies.

Evidence-backed resume workspace tooling for AI-assisted job searches.

MaxOS turns AI agents into persistent, owned, policy-bounded, and auditable worker identities.

Create an AI by chatting - it gets a URL, a memory you can see, a body across your devices, a voice, and a wallet.

SABBA - Security Templates CLI & MCP Server for coding agents that prove every finding by running it

A launch playbook your coding agent can run — take a shippable AI product, agent, SaaS, or devtool from ready to launched, no audience required.

OpenTelemetry SpanExporter for the Vercel AI SDK → Voight. Drop-in observability

Workspace for humans and AI agents. Every block is a doc, a task, a discussion, and agent context — one object, not three copies your agent has to choose between. Context is lexically scoped: pin a rule at the root, every task below inherits it. Self-hosted, MCP-native.

Usage, cost and account toolkit for Claude Code, Codex and other AI coding agents

Native cross-chain swaps: Bitcoin, Ethereum, Solana, Polkadot, Tron, Arbitrum. Quote, swap, track.

The terminal AI agents can drive — and the cockpit to run them from. 99 MCP methods / 21 namespaces, agent Inbox, verified worktree fleets, diff review & rollback. macOS / Linux / Windows, local-first, $0, MIT.

Give your coding agents correct architecture context: design declared once in git, mechanically checked, drift-detected, and rendered into verified briefs

W.A.Y? (Who Are You? : I'm not a developer) — a personal AI harness that learns you, not a framework you learn

AI-human engineering platform where AI agents and humans collaborate as teammates to plan, code, review, and ship in isolated sandboxes. An append-only event store gives a full audit trail and DORA-grade metrics derived from the stream, with built-in financial-compliance templates. FastAPI + LangGraph.

Your AI Quality Agent, putting app testing on autopilot.

Plot anything with zero effort. Editable charts, diagrams, figures, posters, and slides.

ULA – 从收案到文书的律师诉讼全流程 AI 工作台 · WorkBuddy多Expert并行 · 元认知自蒸馏 · 非代码律师友好

Local-first memory for AI-assisted coding. Git shows what changed. Backstory shows why.

Turn untrusted source material into compact, attributed AI skills through evidence gates and deterministic safety checks.

The tested, tool-agnostic AGENTS.md kit — verified templates, a CI linter, and migrators from .cursorrules/CLAUDE.md/Copilot/Windsurf/Cline/Aider.

A ratified constitution + harness-neutral enforcement regime for AI coding agents — hooks, a cold-startable project-state model, and the adversarial-review harness used to harden them.

Immutable, reproducible, least-privilege runtime environments for AI agents

Reference implementations and utilities for SIGN (Sigil Intelligence Graph Notation) — sign-ingest and sign-draft.

Federated MCP gateway: one OAuth 2.1-protected, spec-compliant MCP server in front of all your apps. Backends stay plain HTTP (~120-line contract) or proxied native MCP servers with a per-user token vault. Scope-filtered tools, live discovery, audit trail. Give agents curated, governed access to your whole stack — one URL, one token.

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.

Skills for engineers that care about quality and scalability. Lessons condensed from my blog articles.

Open-source Bittensor subnet for a global AI agent competition in sales intelligence

AI software factory control plane for GitHub issue-driven autonomous workflows, isolated agent sandboxes, governance, artifacts, and human review.

Discover Hackathons, AI Tools, and Developer Opportunities

GitHub-native Loop Engineering skill: evidence-gated Issues and Projects workflows over the official GitHub MCP Server.

Skills de IA para Product Management y Spec-Driven Development, creadas por Javier Garzas (233 Academy) y compartidas con la comunidad.

Public reference body for agent-owned software, co-evolution, and Genesis descent.

AI agent game competition - Reinforcement learning (Monte Carlo Tree Search, Deep Q-learning, Minimax)

Disciplined minimal code orchestrator with gated multi-model review before subagents ever start writing code.

Claude code / Open Code Skill guiding you step by step, to generate continuous-shot video using Gemini web interface

An open-source digital micropad for agentic work — turn your phone or tablet into a physical-feeling control deck for AI coding agents

Open-source Outcome Packs that turn rough ideas into coordinated, independently verified outcomes with Codex.

Unity Editor MCP plugin — AI agents drive Unity via Model Context Protocol. Apache 2.0, ~290 domain-specialized tools including procedural generation, simulation, physics, rendering, and AI behavior from published research papers.

120 brand design systems as plain-text DESIGN.md specs for AI coding agents. Real colors, fonts, components and guardrails, reverse engineered from live production sites.

Plan, execute, and review coding-agent work in one desktop workspace.

The agent-first IDE: run a fleet of coding agents (Claude Code, Codex, Gemini, Copilot, Antigravity) across isolated git worktrees, with a commit-time review gate. macOS · Windows · Linux.

Production-ready memory framework for AI agents. Dual-agent synthesis, self-evolution, composite scoring, multi-tenant isolation. REST API, Python SDK, CLI.

Local, offline safety net for AI coding agents — records every file change and reverts anything (one file → a whole rm -rf), and alerts on secret reads. Claude Code · Codex · Cursor. Zero deps · MIT.

A ready-to-run Buzz workspace in your browser — one command starts the Buzz desktop, relay, storage, and optional Codex, Claude Code, and Goose agents

Privacy-first open-source contract review CLI. Multi-agent AI pipeline strips PII locally (fail-closed) before any cloud call. 21 contract types, 17 AI providers, local-first by default with Ollama. Runs as terminal command, TUI, or Python library.

⇄ Port your session between coding agents (Claude Code, Codex, Grok Build). Type /seshport, open the other agent, resume with full context. On npm, PyPI, crates.io & Homebrew.

Find what your AI agent gets wrong — before you have a rubric. Qualitative eval for PMs.

Cross-session memory and recall for AI agents — git-synced knowledge base, hybrid semantic+TF-IDF search, auto-distillation with secrets scrubbing

A modular Agent Skill system for strategy, creative, content, and brand teams—from brief framing and evidence to insight, narrative, visual direction, and delivery review.

The open-source control plane for AI agents — skill routing, context governance, trustworthy execution, evidence, security, and multi-agent orchestration.

MCP server that drives Google Flow (Veo / Nano Banana) through its internal tRPC API, with a hard credit-cost gate

Multi-Agent Code Evolution Platform — 4 LLMs compete, evolve, and breed code through a JIT compiler judge

MCP server for SERPHouse API. Enables LLMs to perform live, geo-targeted Search engine queries. Built with TypeScript.

Use existing brand guidelines with AI. Turn websites, PDFs, and Figma into a portable .brand runtime through MCP.

Comfy Agent is a tool to use ComfyUI from the CLI. It is suitable for both direct CLI usage and AI-agent-driven automation.

Public source of the Dreadnode capabilities in app.dreadnode.io — agents, tools, skills, MCP servers, and workers.

🧠 ctx is a queryable world model of a codebase, built to ground and govern the language models that modify it

Let AI agents see, debug and drive running WPF apps - visual tree, bindings, screenshots, clicks and text input over MCP. Snoop + Playwright for AI.

Offline-first, precision-first security linter for SKILLS, MCP, plugins, configs and other AI infrastructure.

Multi-LLM content strategy dashboard — research-grounded weekly social calendar for X, LinkedIn, and Reddit with personas, story bank, and performance feedback.

Team operating model for hybrid human+AI delivery: one shared, drift-checked contract for authority, risk tiers, evidence, review, and done — with cloud-expert guardrails (Google Cloud, AWS, Azure, Alibaba Cloud) and Google ADK/Antigravity skills built in, routing agent-craft work to companion skill plugins.

Control plane plugin for Hermes Agent — pool many free/cheap model providers + local AI CLIs, switch brains from Telegram in one command (even while the model is down), delegate work cap-aware with auto-fallback. Run your agent at ~$0 without hitting quota walls.

Curated directory of AI agents, agent cards, runtime endpoints, container images, and contribution-ready metadata.

Local-first action assistant for durable, reviewable software work.

Know what your agents are doing

Run xAI's Grok CLI from any project, with all traffic routed through a VPN (Docker + gluetun, Mullvad by default)