refly-ai/refly

The first open-source agent skills builder. Define skills by vibe workflow, run on Claude Code, Cursor, Codex & more. Build Clawdbot 🦞· APIs for Lovable · Bots for Slack & Lark/Feishu · Skills are infrastructure, not prompts.

What it solves

Refly addresses the reliability gap in AI agents, where "vibe-coded" scripts and black-box logic often lead to production failures. It replaces fragile, one-off prompts with stable, atomic, and versioned "skills"—deterministic agent capabilities that ensure business logic is executed with 100% reliability across different agent frameworks.

How it works

Refly provides a visual IDE and a Model-Native DSL that allows users to codify business logic into executable skills. It features a Copilot-led builder ("Vibe Mode") where users describe intent in natural language to generate skills. These skills are managed in a central registry and executed via a stateful runtime that allows users to pause, audit, and re-steer logic mid-run to prevent hallucinations and ensure compliance.

Who it’s for

It is designed for developers and enterprise teams who need to move beyond simple orchestration tools or heavy code-first SDKs to build governed, portable, and production-ready agent infrastructure.

Highlights

  • Intervenable Runtime: Ability to steer and hot-fix agent logic during execution rather than restarting from scratch.
  • Universal Delivery: Export skills as standard APIs, webhooks (for Slack, Lark/Feishu), or native tools for Claude Code and Cursor.
  • Copilot-led Construction: Rapidly transition from a natural language description of a business process (SOP) to a functional skill in minutes.
  • Extensive Integration: Supports over 3,000 native tools and is fully compatible with the Model Context Protocol (MCP).

Related

  • Project
    AgentEra/Agently

    An AI application runtime framework that provides structured output guarantees, observable action execution, and signal-driven workflows for building reliable AI service backends.

  • Project
    FrancyJGLisboa/agent-skill-creator

    A platform that turns expert human knowledge and business evidence into tested, reusable, and governed AI agent skills for organizational use.

  • Project
    Agent-Field/agentfield

    An open-source control plane that turns AI agent logic into production-ready REST APIs with built-in routing, durable queuing, and observability.

  • Project
    google/agents-cli

    A CLI and skill set that enables coding agents to build, evaluate, and deploy enterprise-grade AI agents on the Gemini Enterprise Agent Platform and Google Cloud.

  • Project
    HoangNguyen0403/agent-skills-standard

    A portable SDLC standards layer for AI coding agents that turns engineering rules into modular, version-controlled skills to ensure consistent code quality while reducing token waste.