Explore the directory

This repository holds the notebooks, exercises and data for a one-day masterclass that teaches engineers and architects to build security and privacy into AI and machine learning products by design rather than adding it afterward.

• Security
• Tutorials

This is a curated archive of agentic AI sample projects and reusable tooling from an AWS-run Agentic AI Acceleration program in Korea, covering production-grade agent architectures built on Amazon Bedrock and AgentCore.

• Tutorials

This project packages Ollama as a serverless worker for the RunPod.io platform, downloading a named model automatically on startup and serving both native Ollama and OpenAI-compatible streaming requests.

• Inference
• Hosting

pawbot is a self-hosted AI agent reachable from a browser, a terminal or chat apps, that reads and writes files, runs commands, searches the web and calls MCP tools, with a Record and Replay feature that replays a past turn offline while the code changes.

• Coding agents
• Debugging

mcp-log-proxy is a small web-based logging proxy that sits between an MCP client and server over stdio, so a developer can watch every message passed between them in a browser while debugging an MCP integration.

• MCP servers
• Debugging

Markdown Vault MCP is a generic MCP server that exposes a folder of Markdown notes, such as an Obsidian vault, to an AI client with hybrid keyword and semantic search, frontmatter-aware filtering, and optional write tools for editing the vault directly.

• MCP servers
• RAG

This project wires the Hermes autonomous agent directly to a local llama.cpp model, and optionally bridges Claude Code through a local LiteLLM proxy, so an agent can plan tasks, call tools and edit files entirely on a user's own hardware with no per-token API cost.

• Agent frameworks
• LLM gateways

Grantex is an open-source delegated authorization protocol for AI agents: it gives each agent a verifiable identity and scoped, revocable authority to act for a human or organization, with an immutable audit trail, complementing OAuth 2.0 and MCP rather than replacing either.

• Security
• Auth

gmlx is a local inference platform for running K-quant and IQ-quant GGUF models on Apple Silicon, serving them over OpenAI- and Anthropic-compatible APIs, with terminal or browser chat, voice input, a local RAG stack and LoRA fine-tuning.

• RAG
• Inference

Ghostlink is a self-hosted, open-source distributed inference fabric that discovers other machines on a LAN and splits a language model's weights across their CPU, GPU and NPU resources, serving the result behind an OpenAI-compatible API with no manual configuration.

• Inference

This repository is a pinned, tested distribution of the upstream Crawl4AI web-crawling toolkit, packaging its native MCP server behind Docker Compose with setup and deployment guidance rather than shipping a custom MCP implementation.

• MCP servers
• Browser automation

This benchmark compares three memory architectures for multi-agent LLM systems - raw conversation history, vector-only RAG, and a structured context graph - on questions that require combining facts from different points in a conversation, using no LLM calls.

• Agent memory
• AI evals

ComfyUI-PromptChain is a prompt-authoring toolkit for ComfyUI that turns the plain prompt box into a structured, code-editor-like workflow with chaining, wildcards and a visual tag-based composer, carrying that structure through posing, generation, viewing and upscaling.

• Generative media

Astromesh is an open-source runtime that standardizes how AI agents execute and interact with external systems: agents are defined declaratively in YAML, routed to different models per role, orchestrated with built-in patterns, and deployed as a REST and WebSocket API.

• Agent frameworks

AsterVec is an embeddable, disk-based vector search engine for on-device AI agent memory and local RAG, keeping only a small navigation structure in RAM so it can search millions of embeddings within a few hundred megabytes of memory.

• Agent memory
• Databases

Agent News is a free, unauthenticated MCP server that gives an AI agent access to a curated feed of verified, cryptographically signed news articles about the AI-agent and MCP ecosystem, each tagged with a confidence level and a provenance signature.

• MCP servers

Ablation is a reverse engineering framework offering the same core disassembly, decompilation and binary analysis capabilities as tools such as Ghidra, IDA Pro and Binary Ninja, that becomes a fully autonomous reverse engineering tool when paired with Claude Code or OpenAI Codex.

• Debugging
• Security

A Python and SymPy-based symbolic code-generation toolkit for numerical relativity and relativistic astrophysics that turns symbolic physics equations into standalone C/CUDA applications, Einstein Toolkit thorns, Charm++ projects and JAX/Python workflows.

• Libraries

A developer tool for parsing Mellanox hardware and software steering dumps, covering both the HWS and SWS acceleration paths.

• Debugging
• Networking

A cross-repository control project integrating Newton-Muon and Locoprop-style preconditioned optimizers, with FSDP and MuP support, into Megatron-LM and TransformerEngine for training mixture-of-experts models at scale.

• Models

A multi-index, multi-vector, document-based vector database and search engine with a JSON RESTful API, combining vector, keyword, geo and filter search in a single query.

• Databases

A face and fingerprint authentication system for Linux login, lock screen, sudo and app prompts, using infrared cameras with anti-spoofing checks on by default and TPM-sealed embeddings.

• Security
• Auth

Roku Dev Studio is a cross-platform desktop application for Roku channel developers that wraps remote control, app sideloading, a debug console, BrightScript debugging, network inspection and an MCP server for AI coding agents into one GUI and one CLI.

• MCP servers
• Debugging

QodeX is a local-first agentic coding CLI that defaults to running a local model through Ollama or LM Studio, edits and tests code with deterministic guardrails around the model, and can run its shell inside a Docker sandbox instead of directly on the host.

• Coding agents