The self-improving all channels AI agent. Self-healing. Fully autonomous. Single Rust binary.

Server-enforced workflow discipline for AI agents. An MCP server providing persistent work items, dependency graphs, quality gates, and actor attribution. Schemas define what agents must produce — the server blocks the call if they don't. Works with any MCP-compatible client.

若爱 (IfAI) - 基于 Tauri 2.0 构建的跨平台 AI 代码编辑器 | A Cross-Platform AI Code Editor Built with Tauri 2.0

Open-source Environment toolkit of claw-like agents, support task/harness generation and evaluation

A light-weight and straightforward system for spec-driven development with Claude Code or OpenAI Codex

A minimal, composable AI agent harness framework in Python.

Open-source framework for agent-native software. Rails for apps that build themselves - with a copilot inside.

AI-powered development framework with task management, 41 agents, 83 skills, and MCP tools for Cursor, Claude Code, Gemini, Codex & OpenCode. File-based memory that survives across sessions.

Ostrom-compliant commons governance & development system for AI agents to assist building products from requirements to production & deployment.

Let's Qitos! A torch-like agent-native framework for researchers.

Your plugins, skills, MCP servers, hooks, conventions, and governance packaged into a single config

AMP (Agent Mailer Protocol) is an async mailbox system for AI agents. Instead of chaining tools, writing DAGs, or building RPC glue — agents just send messages to each other. Works with Claude Code, Codex, Cursor, OpenClaw and any custom agents. Designed for real-world, long-running, iterative coding workflows

Compile CLAUDE.md into a lockfile, then enforce repo changes against it in local runs, staged diffs, or CI.

A skill-driven Harness Engineering framework

Reward model engineering harness for evolutionary rubric search, deployable RM artifacts, online scoring, and RL experiment lineage.

Fabricatio is a Python library designed for building applications that leverage Large Language Models (LLMs) using an event-based agent structure. It utilizes Handlebars for templating and provides a robust framework for managing tasks, workflows, and toolboxes, enabling developers to create intelligent and efficient applications.

Racker is an experimental harness tool for provisioning and running operating system containers.

An architecture-level AI harness that compiles constraints and NFRs into explicit, reviewable architecture decisions. Combines a deterministic compiler, curated pattern registry, and agent workflow skills to drive approval, re-approval, and implementation against an architectural contract.

Unified AI agent configuration management for Codex, Claude, and Copilot by wild

A compact, schema-first standard for the boundary between cognition and execution.