Explore the directory

A reference collection documenting how to configure OpenAI's Codex CLI: its commands, subagents, skills, plugins, marketplace, memories and MCP server support.

• Tutorials

A native macOS app that wraps a self-hosted vMLX inference server, serving language, vision and image-generation models over an OpenAI-compatible API on Apple Silicon with no Python setup.

• Inference

A TypeScript runtime that turns one-off agent chats into auditable, replayable DAG workflows for a homelab or home server, driven primarily by voice.

• Agent frameworks
• Workflow automation

An autonomous penetration-testing agent that drives an OpenAI-compatible LLM gateway against a Kali Linux Docker sandbox, running a structured recon-analysis-exploit-report pipeline through a real-time terminal interface.

• Security

A hardware-focused fork of vLLM tuned for dual RTX 2080 Ti and other Turing (SM75) GPUs, adding launcher profiles and validated inference routes for running 27B to 35B class models locally.

• Inference

A Markdown-only workflow plugin for AI coding agents that turns large software changes into a spec-driven loop of project analysis, task decomposition, GitHub issue and PR tracking, and adaptive control.

• Agent skills

A PyTorch-native distributed training library for fine-tuning large language and vision-language models directly from their Hugging Face checkpoints, with recipes for many current open-weight model families.

• Models

A vision-language-model-based retrieval-augmented generation pipeline that embeds documents as images instead of parsed text, and its newer EVisRAG variant that reasons over per-image evidence for multi-image questions.

• RAG

An implementation of Nvidia's Parakeet automatic speech recognition models for Apple Silicon using MLX, with a command-line transcription tool.

• Speech

A native cross-platform app that gives every local AI coding agent one command center, running each agent in its own tmux pane over an end-to-end encrypted peer-to-peer or relayed connection between machines.

• Coding agents

A reinforcement-learning framework for post-training unified multimodal models, coordinating rollout generation, scoring and policy updates across pluggable trainers, rollout engines and algorithms.

• Models

A collection of Model Context Protocol servers from Neo4j Labs that let an MCP client such as Claude Desktop or an IDE assistant query a Neo4j database and the Aura cloud service in natural language.

• MCP servers
• Databases

A ComfyUI custom node that adds one-click virtual VRAM and multi-GPU distribution for UNet, CLIP and VAE loaders, offloading model layers to other devices to free VRAM for larger generations.

• Inference
• Generative media

A GitHub Action that installs and runs Claude Code inside a workflow so a team can build custom automation on top of it; it enforces no trust boundary of its own, so untrusted-input workflows should use the separate claude-code-action instead.

• Coding agents
• CI/CD

A modular runtime that orchestrates a marketplace of plugins, agents and skills into structured pipelines for AI-assisted software development, installed for Claude Code, Codex, OpenCode, Cursor and Kiro.

• Coding agents
• Agent frameworks

A runtime built on AgentField that spins up an autonomous team of planning, coding, review and testing agents from a single API call to produce production-grade pull requests.

• Coding agents
• Agent frameworks

A pair of Agent Skills packages that teach coding agents about building with Google's Gemma models and about training or fine-tuning them locally, installable through the Vercel or Context7 skills CLIs.

• Agent skills
• Models

A set of over 30 hands-on Jupyter notebook labs that show how to secure, meter and observe large language models, MCP servers and AI agents behind Azure API Management's AI Gateway.

• LLM gateways
• Tutorials

FanBox is a macOS desktop app putting a local file browser, live file previews and a real embedded terminal in one window, so a developer can find scattered agent-created projects, run Claude Code or Codex, and watch exactly which files it changes.

• Coding agents

Base LLM is a full-stack Chinese-language course going from traditional NLP foundations, word vectors and RNNs, through the Transformer architecture, to modern LLMs, covering hand-written BERT, GPT and Llama 2, parameter-efficient fine-tuning, RLHF and deployment.

• Tutorials

CodeAlmanac is a living wiki for a codebase, written and kept current by AI coding agents, that captures the decisions, invariants, past failures and cross-file workflows source code alone cannot express, stored as Markdown and reviewed through Git like any change.

• Agent memory
• Documentation tools

The Alpaca MCP Server is Alpaca's official MCP server for its brokerage Trading API, letting an AI assistant such as Claude, Cursor or VS Code place stock, option and crypto trades, manage a portfolio and pull real-time market data from plain-language requests.

• MCP servers
• APIs

An enterprise-grade AI API gateway, built as an extended fork of one-api, that adds subscription plans, active-active clustering, and precise billing for reselling or managing access to multiple model providers.

• LLM gateways

This is a setup guide combining an Obsidian vault as Claude Code's persistent, Zettelkasten-style memory across sessions with a companion tool, Graphify, that maps a codebase into a knowledge graph, so Claude Code stops re-reading its files every session.

• Agent memory
• Tutorials