Meta-Harness Engineering (MHE) is a Python reference runtime for the Meta-Harness, which is a research and engineering runtime rather than a production orchestrator.
PMake is a meta-build system that translates high-level project manifests into executable automation for agent assistants.
A standalone Python package for interacting with Qwen3 models via Dashscope API.
A Python package that provides semantic search capabilities using embeddings and vector stores.
A powerful command execution tool with automatic logging and comprehensive history management.
This repository collects reusable prompt templates and worked examples for scheduling multiple AI agents to collaborate in parallel.
A comprehensive, production-ready system for converting C# codebases to idiomatic, performant C++ code.
A high-performance Julia implementation of the Gridpoint Statistical Interpolation (GSI) core analysis algorithms, ported from the GSI/EnKF Fortran codebase.
A Python implementation of the Local Extreme Learning Machine (locELM) method with domain decomposition for solving linear and nonlinear partial differential equations.
This document summarizes the complete rewrite of the Claude Code Development Guide
Agent-Society envisions a digital civilisation where large language model (LLM) agents collaborate, compete, and evolve under spatial (compute and memory) constraints rather than biological time.
A unified code flow analysis and performance instrumentation tool for C++ and Fortran applications, with specialized support for domestic supercomputing platforms.
Local-first agent space (QQ Space-like) for posts, comments, reactions and footprints in the terminal and web.