Create Unlimited OpenClaw Skills | Full Tutorial (Upstage Studio Skill)

Key Concepts

• Open Claw: An AI agent framework that gains functionality through "skills."
• Skills: Modular instructions or scripts (Python) that define how an agent performs specific tasks.
• Workspace: A designated directory where Open Claw stores files, scripts, and skill definitions for transparency and control.
• Upstage Studio: An AI-powered document processing platform used to parse, classify, and extract structured data (JSON) from invoices.
• Gateway: The core service layer of Open Claw that must be restarted to apply configuration changes.

1. Setting Up Open Claw

To begin, the user installs Open Claw via the CLI (compatible with Mac, Linux, and Windows).

• Configuration: Upon first launch, the user must configure the LLM provider (e.g., OpenAI) and provide an API key.
• Workspace Management: To maintain control, the user sets a specific local directory as the default workspace using the command: Open Claw config set agents default workspace [path].
• Gateway Restart: After any configuration change, the command Open Claw gateway restart is required to initialize the new settings.

2. Creating Skills: Methodology

A "skill" in Open Claw is essentially a set of instructions that tells the agent how to execute a specific tool. There are two primary ways to create them:

Manual Creation

1. Directory Structure: Create a skills folder within the workspace.
2. Skill Definition: Create a sub-folder (e.g., stock_price) containing a skill.md file. This file acts as the instruction manual for the agent.
3. Tool Integration: Place the executable Python script (tools.py) in a scripts folder. This script contains the logic (e.g., calling an external API to fetch stock data).
4. Verification: Use Open Claw skills list to confirm the agent recognizes the new skill.

Automated Creation

The user can prompt Open Claw to create a skill by providing the necessary logic or API instructions directly in the chat interface. The agent then verifies if the skill exists and automatically generates the required files in the workspace.

3. Real-World Application: Invoice Processing

The video demonstrates automating the extraction of data from multiple PDF invoices using a combination of Open Claw and Upstage Studio.

• Workflow:
  1. Upstage Studio Setup: Upload an invoice to Upstage Studio to define the extraction schema.
  2. Parsing & Extraction: The tool parses the PDF and extracts key fields (e.g., invoice number, due date) into a structured JSON format.
  3. Integration: The API endpoint from Upstage is integrated into an Open Claw skill.
  4. Batch Processing: The user instructs Open Claw to process a folder of invoices. The agent automatically generates a script (process_invoices) to iterate through the files, call the Upstage API, and save the results as structured JSON files in a designated output folder.

4. Key Arguments and Perspectives

• Modularity: The presenter argues that Open Claw’s power lies not in the base model, but in the ability to add custom skills. This allows the agent to handle specialized tasks that general-purpose models might struggle with.
• Transparency: By setting a custom workspace, the user maintains full visibility and approval authority over the code the agent generates and executes.
• Efficiency: Using specialized tools like Upstage Studio for document parsing is presented as superior to relying solely on an LLM, as it ensures higher accuracy for structured data extraction.

5. Notable Quotes

• "Open Claw is powerful, but if it doesn't know what to do, then nothing to do much with this. That's when we have skills."
• "Skill is nothing but an instruction on how to run the tool. Or it could be a general instruction on what the agent should do."

6. Synthesis

The video provides a practical framework for extending AI agent capabilities. By combining the Open Claw framework with external specialized tools like Upstage Studio, users can transition from simple chat-based interactions to complex, automated workflows. The core takeaway is that by defining clear, modular "skills," users can transform a generic AI agent into a highly efficient, task-specific automation engine capable of handling batch data processing and structured information extraction.