A sophisticated reasoning framework that orchestrates AI agents using a dual-system approach for both rapid execution and deep strategic planning.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install yuyonghao-react-orchestrator
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install yuyonghao-react-orchestrator using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is React Orchestrator?
React Orchestrator is a powerful reasoning framework designed for the next generation of Openclaw Skills. It implements a dual-system architecture inspired by cognitive science: System 1 for fast, intuitive execution (ideal for simple queries and factual retrieval) and System 2 for deep, reflective reasoning (perfect for complex planning and multi-step strategy). \n\nBy leveraging the ReAct (Reason + Act) loop combined with a Reflexion mechanism, this skill allows AI agents to not only act but to think, observe, and self-correct their path in real-time. This ensures that developers can build more reliable and autonomous agents that intelligently manage their own computational resources based on task complexity.
React Orchestrator Use Cases
- Handling rapid factual queries like weather reports or simple file reads via System 1.
- Executing complex multi-step tasks such as full project generation or deep industry research via System 2.
- Creating self-correcting agents that use reflection to identify and fix reasoning errors during execution.
- Managing high-density tool environments where intelligent routing is required to maintain performance within Openclaw Skills.
How React Orchestrator Works
- The orchestrator receives a user query and immediately evaluates its complexity based on semantic markers.
- It routes the task to System 1 (3-5s execution) if simple, or System 2 (10-15s execution) if deep reasoning is required.
- The ReAct cycle begins, where the agent generates a 'Reasoning' step, followed by an 'Action' using registered tools.
- After each action, the agent 'Observes' the result and decides whether to iterate or finalize the answer.
- For System 2 tasks, a Reflexion step occurs periodically to audit the current plan and optimize the strategy for the remaining iterations.
- The final output is delivered along with a comprehensive execution history for debugging and transparency.
React Orchestrator Setup
To integrate this framework into your project, use the following commands:
# Install via ClawHub
clawhub install react-orchestrator
# Or manual installation
git clone https://github.com/YOUR_GITHUB/skills/react-orchestrator
cd skills/react-orchestrator
npm install
React Orchestrator Data Schema & Taxonomy
The skill manages execution data and tool registries using the following structure:
| Property | Type | Description |
|---|---|---|
answer |
String | The final synthesized response from the agent. |
history |
Array | A complete log of thoughts, actions, and observations. |
mode |
String | Indicates if System 1 or System 2 was utilized. |
iterations |
Number | Total number of ReAct cycles performed. |
duration |
Number | Total time taken in seconds. |
complexityScore |
Number | The calculated difficulty of the task (0.0 to 1.0). |
React Orchestrator Advanced Features
- Intelligent Complexity Assessment: Automatically detects task difficulty to save tokens and time.
- Reflexion Mechanism: Periodic self-audit steps that allow the agent to pivot if a chosen strategy is failing.
- Semantic Tool Matching: Uses keyword and semantic analysis to find the most relevant tools in your Openclaw Skills library.
- Timeout Protection: Built-in safety limits (30s for System 1, 120s for System 2) to prevent runaway agent loops.
- Verbose Logging: Detailed execution traces for developers to audit the decision-making process of their AI agents.
SKILL.md
Loading
Related Openclaw Skills
Unreal Engine Integration for OpenClaw
A comprehensive integration suite for connecting OpenClaw agents to Unreal Engine 5.x to automate editor tasks, scaffold C++ plugins, and drive Blueprint-driven workflows.
Ron: Skeptical Reviewer Agent
Ron is a skeptical reviewer agent that scrutinizes code, reasoning, and decisions to identify flaws before they are finalized.
Auto Optimizer
A built-in performance management suite for monitoring, analyzing, and automatically optimizing the execution of AI coding agent skills.
Code Refactor Skill
An intelligent code refactoring assistant that analyzes code quality and automatically applies structural improvements to reduce technical debt.
PowerShell Sandbox
A restricted execution environment for PowerShell scripts that enforces command whitelisting, resource limits, and timeout protection.
OpenClaw SaaS Builder
A comprehensive orchestration skill for building SaaS projects end-to-end using OpenClaw as the driver and Gemini CLI as a planning copilot.
Featured*
