AgentFlow uses Python’s graph-based DSL to orchestrate AI coding agents with parallelism, iteration, and remote execution. It supports Codex, Claude, and others.
AutoSkill is a Python framework enabling LLM agents to extract, version, and evolve skills from dialogues, providing a persistent long-term memory system for AI agents.
Clawd on Desk uses an Electron-based desktop pet with pixel-art animations to visualize AI coding agents’ activity in real time, enabling glanceable observability and inline permission handling.
ClawMetry offers a zero-configuration real-time dashboard for OpenClaw AI agents, visualizing message flows, token usage, and operational alerts with easy setup.
Clay unifies local repos into a browser-based dashboard with persistent AI teammates that remember across sessions and iterate autonomously via overnight coding loops. Self-hosted with no cloud lock-in.
homeassistant-ai/skills provides a modular collection of Home Assistant skills for AI coding agents supporting the Agent Skills standard, easing integration with tools like Claude Code and Copilot.
NarratorAI CLI Skill uses a machine-readable SKILL.md file to enable AI agents to automate movie narration video production via a CLI tool. Supports multiple agents and rich media resources.
opencode-power-pack translates Anthropic’s Claude Code plugins into OpenCode SKILL.md files, enabling multi-agent AI coding workflows across platforms. Here’s how it works under the hood.
gtm-agents bundles 67 GTM plugins, 92 AI agents, 52 business skills, and 20 workflow orchestrators for Claude Code, saving 15+ hours/week on sales and marketing busywork.
red-run orchestrates Claude Code AI agent teams across the full pentest kill chain using persistent teammates and semantic routing. Explore its architecture, strengths, and quickstart.
The google-docs-mcp server exposes Google Docs, Sheets, Drive, Gmail, and Calendar as callable tools for AI agents like Claude Desktop, enabling deep document and email automation via OAuth2.
LLM4Pentest aggregates 40+ research papers and tools tracking the evolving role of LLMs in automated penetration testing, highlighting progress and limitations.
mcp-selenium exposes Selenium WebDriver as typed MCP tools for AI agents, supporting multi-browser automation with WebDriver BiDi diagnostics. Setup is easy with npx commands.
native-devtools-mcp provides a Rust-based MCP server letting AI agents control native desktop apps on macOS/Windows, Chrome/Electron browsers, and Android devices. It features macOS AX dispatch for precise, focus-free app control.
opcode is a cross-platform Tauri desktop app that wraps Claude Code CLI with a GUI, session checkpointing, custom AI agents, MCP server management, and usage analytics.
OpenClaw Agents deploys paired AI agents using shell scripts for adversarial collaboration in OpenClaw gateway, supporting multi-channel or local workflows with safe config merges.
OpenUI provides a TypeScript CLI and browser UI to spawn, organize, and monitor multiple AI agents with session persistence and real-time status. Easy to start and extend.
Ori-Mnemos provides a TypeScript CLI tool for managing AI agent memory with scoped vaults and activation modes, enabling persistent and flexible context handling.
SceneSmith uses GPT-5-powered agents to generate physically plausible 3D indoor scenes from text prompts, ready for robotics simulation without manual cleanup.
Explore a detailed 17-phase AI engineering roadmap for 2026, focusing on multi-LLM orchestration, RAG, AI agents, and production-ready skills with 51 hands-on projects.
