#harness-engineering Repositories
Discover 64 active open source GitHub repositories tagged with #harness-engineering, accumulating 353,468 total stars (averaging 5,522 stars per project). Leading this topic is infiniflow/ragflow with 94,501 stars.
65 repositories total
RAGFlow is an agent-native database: a Retrieval-Augmented Generation (RAG) engine that grew into context infrastructure.
Your Personal AI Assistant; easy to install, deploy on your own machine or on the cloud; supports multiple chat apps with easily extensible capabilities.
How Python does AI. Agents, realtime voice, image generation, embeddings. Every model, every interface, typed end to end.
DeepCode: Open Agentic Coding (Agent Harness & Loop Engineering & Multi-Agent Orchestration)
A meta-skill that designs domain-specific agent teams, defines specialized agents, and generates the skills they use.
The open-source agent harness - the runtime layer that turns an LLM into a working agent.
Nexent is a zero-code platform for auto-generating production-grade AI agents using Harness Engineering principles โ unified tools, skills, memory, and orchestration with built-in constraints, feedback loops, and control planes.
Awesome list for AI agent harness engineering: tools, patterns, evals, memory, MCP, permissions, observability, and orchestration.
The first vision plugin for DeepSeek Harness, and the vision bridge for every text-only coding agent. Paste an image, get structured JSON evidence (OCR, layout, semantics).
HarnessRouter Community Edition: the self-hosted, Apache-2.0 edition of the unified interface for agent harnesses. Run Codex, Claude Code, Hermes, PI, DSH, and more through one API, with sessions, streaming, files, cancellation, and failure handling. Implements the Unified Harness Protocol (UHP), an open standard. Your keys, your infrastructure.
GoClaw - GoClaw is OpenClaw rebuilt in Go โ with multi-tenant isolation, 5-layer security, and native concurrency. Deploy AI agent teams at scale without compromising on safety.
QuantMind is an open source agent-native knowledge extraction and retrieval framework for quantitative finance.
The Context Layer for unstructured data: typed, versioned datasets over S3, GCS, Azure
A model-driven harness for AI coding agents: semantic modeling, app generation, progressive API Assist, and evidence-based verification.
Surge-style whole-home gateway and control plane for macOS with IPv4/IPv6 supportโ mihomo TUN, dnsmasq-powered DHCP/DNS, per-device routing, and an agent-friendly validation workspace.
DeepBot is a system-level AI assistant built for both personal productivity and enterprise workflows โ one-click setup, seamless experience, and native Feishu integration.
An open-source Harness Engineering platform for coding agentsโdefine harnesses as code, run controlled experiments, inspect evidence, and compare outcomes. Turn task evidence into actionable team and organization insights.
Multi-agent graph harness for Claude Code, Codex, and other coding agents. Separate agents review and test the work. Runs locally on your subscription.
Plan, implement, and verify software with your favorite coding agents. Work in the app or browser, or in your own terminal, IDE, or agent app.
An awesome list of Agent Harness engineering resources, including GitHub projects, tools, benchmarks, and practical guides.
๐ Ranked list of 187 AI agent harnesses, plus templates, playbooks, MCP, and learning resources. Rescored weekly.
The skillbook for recursive self-improvement (RSI). NVIDIA's latest proof, unpacked with timeless engineering. Distilled from the OpenAI harness engineering hackathon winner's zero-touch method.
AgentSociety 2 is a modern, LLM-native agent simulation platform designed for social science research and experimental design. It provides a flexible framework for creating and managing intelligent agents in simulated environments.
Tianshu Harness is a terminal programming agent. The core question it is designed to answer is: How can a model deliver reliablyโwith no goal drift, completion backed by evidence, and verification forming a closed loop, rather than saying โIt should be fixed nowโ? To achieve this, it establishes a cognitive execution environment (CVM) between the model and the real world, externalizing goals, state, evidence, resources, permissions, and termination conditions from the conversation history for continuous runtime management. It is also highly optimized for DeepSeek V4.
The cloud infrastructure behind puppyone. Hosts projects as Git repositories for your team and agents.
Turn any repo into an agent-ready workspace for Claude Code, Codex, Cursor, and other coding agents.
Design a better vision toolkit and skills for text-only models to "see" images, supporting multi-image understanding, image Q&A, frontend UI restoration, GUI automation, etc., with optional seamless integration with multiple mainstream agents, and direct recognition of pasted images | A vision toolkit and skill designed for text-only LLMs โ image Q&A, long-screenshot OCR, frontend UI restoration, and GUI automation, with optional seamless integration for Codex, Claude Code, Pi, Oh My Pi, and OpenCode
Learn harness engineering by building Claude Code from scratch. Free, open-source course: 8 articles, 6 videos, 1 codebase.
OpenLoomi is an open-source AI coworker. It connects your work tools, understands what youโre working on, and tells you what needs your attention, why it matters, and what to do next. Itโs your open-source Attention Agent.
๐ฅ The strongest free web search plugin for DeepSeek Harness, and the search bridge for every model without native web access. Free, no signup, no API key. Ask the web or X, get structured JSON evidence. | ๐ฅ The strongest free web search plugin for DeepSeek Harness: free, no signup, no API key. Add search capabilities to models without web accessโquery the web or X and get structured JSON evidence (search, scraping, citations).
Attention is your scarcest resource. Chief is the local-first layer that guards it โ turning every agent, alert, and feed into one honest call: interrupt, or not.
๐ Dante Cloud is a cloud-native microservice foundation that supports both blocking and reactive services running in parallel, as well as an intelligent foundation for AI Agent development. Designed based on Domain-Driven Design (DDD), it centers on high-quality code and a low number of security vulnerabilities. It is highly modular and componentized, covering a multi-level security system including IoT device authentication, Chinaโs Classified Protection Level 3, and SM cryptographic digital envelope encryption for interfaces, while also providing a multi-tenant solution. It uniquely enables flexible switching between microservice and monolithic architectures with a single codebase, empowering the development of enterprise-grade intelligent applications. ๐๐Give it a star and stay tuned for updates!
Not a Q&A bot, but a digital double that can get real work done. Share agents and project context so all project members can collaborate. At the same time, agents learn from documents and work, abstracting and summarizing their experience into the project context just like people do.
AgentChat is an LLM-based intelligent agent communication platform, built-in with default Agents and supporting user-defined Agents. Through multi-round conversations and task collaboration, Agents can understand and assist in completing complex tasks. The project integrates technologies such as LangChain, Function Call, MCP protocol, RAG, Memory, HITL, Skill, Milvus, and ElasticSearch to achieve efficient knowledge retrieval and tool calling, using FastAPI to build high-performance backend services.
A copy-paste system prompt that has your coding agent (Claude Code, Codex, Cursor) build a self-improving agent system for any computer work. Try the lead follow-up demo at the website link.
AI workflow automation plugin for intelligent code generation with Claude/Codex
KWeaver Core is a harness-first foundation for enterprise decision agents. It turns fragmented data, knowledge, tools, and policies into governed context, safe execution, and verifiable feedback loops. With semantic modeling, real-time access, runtime control, and TraceAI, it helps AI systems reason, adapt, and act reliable in complex enterprises.
Open-source agentic data engineering harness for dbt, SQL, and cloud warehouses. 100+ tools, 10 warehouses, AI-powered.
Enterprise Agent Harness & MCP Gateway: connect APIs, databases, file servers and 600+ connectors (SAP, Odoo, Etsyโฆ) to Claude, ChatGPT, Microsoft Copilot and your agents. Memory, skills, knowledge graph, SSO/RBAC, audit log and governance built in. Your AI company brain.
Frequently Asked Questions
What is the most popular #harness-engineering repository on GitHub?
The top ranked open source #harness-engineering project on GitHub is infiniflow/ragflow with 94,501 stars. RAGFlow is an agent-native database: a Retrieval-Augmented Generation (RAG) engine that grew into context infrastructure.
Which programming languages are most used in #harness-engineering repositories?
Among the top ranked #harness-engineering projects on GitHub, the most common programming languages are Python, TypeScript, Go, JavaScript.
How is the #harness-engineering repository ranking calculated?
We track 64 active open source repositories tagged with #harness-engineering (500+ stars), totaling 353,468 stars with an average of 5,522 stars per repository. Rankings update daily from the GitHub API.