An intelligent task decomposition and model orchestration engine that transforms complex prompts into manageable, model-optimized execution steps.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install workflow-decomposer
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 workflow-decomposer 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 Workflow Decomposer?
The Workflow Decomposer is a high-level orchestration skill designed to handle sophisticated AI tasks by breaking them down into granular, executable steps. It serves as a strategic layer for Openclaw Skills, ensuring that every part of a complex request is assigned to the most appropriate LLM based on specific reasoning, coding, or creative requirements.
By implementing a structured reasoning model priority system, it optimizes resource allocation and provides full transparency through detailed progress tracking. Whether you are building a full-stack application or conducting deep data analysis, this skill manages the lifecycle of the task from initial decomposition to final delivery, including automated troubleshooting for stuck processes.
Workflow Decomposer Use Cases
- Orchestrating multi-step software development projects where different models handle architecture, coding, and testing.
- Managing long-running research tasks that require distinct phases of scanning, analysis, and report generation.
- Diagnosing and resolving bottlenecks in AI workflows by providing alternative execution paths.
- Tracking real-time progress of complex agentic operations to ensure visibility for the end user.
How Workflow Decomposer Works
- Task Acquisition: The skill receives a high-level objective and identifies its core requirements and complexity scope.
- Strategic Analysis: It determines the necessary skills (coding, analysis, writing) and identifies dependencies between different task components.
- Step Decomposition: The master task is split into manageable sub-tasks, each with estimated timeframes and clear success criteria.
- Model Matching: Each sub-task is assigned to a specific model based on a priority hierarchy, favoring the strongest available reasoning models.
- Execution and Tracking: Steps are executed in sequence, with the skill generating detailed reports on model usage and current progress (e.g., Step 3/8).
- Problem Resolution: If a step exceeds its time limit or fails, the skill diagnoses the error and suggests at least two alternative solutions.
Workflow Decomposer Setup
To deploy this component within your Openclaw Skills setup, follow these installation steps:
# Navigate to your skills directory
cd openclaw-skills/workflow-decomposer
# Ensure dependencies for the optional progress tracker are met
pip install -r requirements.txt
# Configure your model capability references
cp references/model-capabilities.example.md references/model-capabilities.md
Ensure that your environment variables are configured to allow access to the specified reasoning models (e.g., Qwen series, Claude Code).
Workflow Decomposer Data Schema & Taxonomy
The Workflow Decomposer maintains a structured taxonomy for every task it processes, ensuring data consistency across the Openclaw Skills framework:
| Metadata Field | Description |
|---|---|
| Decomposition Model | The specific model used to perform the task breakdown. |
| Task Overview | A concise summary of the primary goal. |
| Progress Matrix | A table mapping steps, descriptions, assigned models, and execution status. |
| Step Detail | Granular inputs, outputs, and optimized prompts for the assigned model. |
| Diagnostics | Log of errors, including logic failures, timeouts, and resource unavailability. |
Workflow Decomposer Advanced Features
- Model-Specific Prompt Optimization: Generates distinct instruction formats tailored for Reasoning, Fast, Code, or Creative LLMs.
- Priority-Based Orchestration: Implements a tiered logic that prioritizes the latest reasoning models and Qwen-series LLMs for complex logic.
- Dynamic Workflow Adjustment: Real-time restructuring of remaining steps based on the outcome of previous sub-tasks.
- Automated Multi-Path Troubleshooting: Provides 'Plan B' and 'Plan C' scenarios when workflows encounter technical or logical hurdles.
SKILL.md
Loading
Related Openclaw Skills
ROS Skill for Robotics Control
A powerful interface for AI agents to monitor and control ROS and ROS2 robotic systems using rosbridge WebSocket communication.
Crypto Price by LPDawn
A real-time cryptocurrency price tracker that fetches live market data from Binance using the ccxt library.
YouTube Transcript
A high-performance skill to programmatically extract transcripts and subtitles from any YouTube video with timestamp precision.
Traffic Cam
A specialized skill for locating real-time traffic camera feeds and live highway broadcasts to monitor road conditions visually.
Common-Fetcher
A unified data collection framework that provides AI agents with robust multi-source information gathering, AI-driven scoring, and summarization capabilities.
ResearchVault
An autonomous state manager and orchestration engine for persistent, high-velocity AI-driven research.
Featured*
