Every card below was actually executed by the lab — under-the-radar repos that installed clean and did what they claim, verified in the sandbox, not guessed from the README. From 950 repos tested so far.
Hound
Hound is an AI-powered security auditing tool that builds and refines adaptive knowledge graphs to perform deep code reasoning.
Insight Installed cleanly on the first try; its own test suite ran — 92 tests passed; the demo actually ran and produced real output.
github.com/scabench-org/hound ↗
Smart-Thinking MCP Server
A Model Context Protocol (MCP) server that provides local, graph-based multi-step reasoning and memory management.
Insight Installed cleanly on the first try; its own test suite ran — 20 tests passed.
github.com/Leghis/Smart-Thinking ↗
Gemini Fullstack LangGraph Quickstart
A full-stack application demonstrating a research-oriented AI agent built with LangGraph and Google Gemini.
Insight The project contains a complete codebase with a frontend, backend, and deployment configurations.
github.com/google-gemini/gemini-fullstack-langgraph-quickstart ↗
Fooocus
Fooocus is a streamlined interface and implementation for Stable Diffusion image generation.
Insight The project has a complete file structure, including Dockerfiles, configuration files, and a comprehensive set of Python scripts, indicating a mature and released application.
github.com/lllyasviel/Fooocus ↗
Video-o3
Video-o3 is a framework for long video multi-hop reasoning that mimics human behavior by actively exploring video content.
Insight Video-o3 is a framework for long video multi-hop reasoning that mimics human behavior by actively exploring video content.
github.com/MCG-NJU/Video-o3 ↗
Deep RAG
Deep RAG is an advanced Retrieval-Augmented Generation system that enhances AI understanding by preserving file structures and generating a 'knowledge map' instead of jus.
Insight Installed cleanly on the first try.
github.com/boluo2077/deep-rag ↗