OpenSkills enables running Claude Skills locally using any LLM. It benefits operations teams by allowing customization and control over AI workflows. It connects to various LLMs like Ollama and Qwen-CLI, integrating into existing automation pipelines.
git clone https://github.com/instavm/open-skills.gitOpenSkills enables you to execute Claude Skills entirely on your local machine using a sandboxed container environment on Mac. Rather than uploading sensitive files to cloud services, you can process documents, spreadsheets, presentations, and images locally while maintaining complete data privacy. The tool integrates with MCP-compatible AI tools including Claude Desktop, Gemini CLI, Qwen CLI, and OpenAI agents, allowing you to leverage specialized skills for file processing tasks. You can import official skills from Anthropic's repository or create custom skills to extend functionality. Setup involves configuring your AI tool to connect to the local MCP server and installing required dependencies.
1. Install OpenSkills and the required LLM (e.g., Ollama or Qwen-CLI) on your local machine. 2. Prepare your input data and define the skill you want to run. 3. Use the prompt template to specify the skill name, LLM, input data, and validation criteria. 4. Run the skill locally and review the output. 5. Validate the results against your criteria and make adjustments as needed.
Extract tables from multiple PDFs and combine into a single Excel spreadsheet
Create professional PowerPoint presentations from markdown outlines
Batch process images: crop to specific aspect ratios and rotate
Fill out PDF forms with data from CSV files
No install command available. Check the GitHub repository for manual installation instructions.
git clone https://github.com/instavm/open-skillsCopy the install command above and run it in your terminal.
Launch Claude Code, Cursor, or your preferred AI coding agent.
Use the prompt template or examples below to test the skill.
Adapt the skill to your specific use case and workflow.
Run the [SKILL_NAME] skill locally using [LLM_NAME] with the following input: [INPUT_DATA]. Ensure the output follows the expected format and validate the results against [VALIDATION_CRITERIA].
Skill Name: Data Extraction Input Data: "Extract customer names and email addresses from the following text: 'Contact us at [email protected] or [email protected]. Our team includes John Doe and Jane Smith.'" Output: {"customer_names": ["John Doe", "Jane Smith"], "email_addresses": ["[email protected]", "[email protected]"]} Validation Criteria: Ensure all extracted names and emails are correctly identified and formatted. The skill successfully extracted the customer names and email addresses from the provided text. The output format is correct, and all data meets the validation criteria.
24/7 AI-powered support ticket resolution
Run large language models locally on your machine
AI assistant built for thoughtful, nuanced conversation
Get more done every day with Microsoft Teams – powered by AI
Agentic AI Workflow platform
Connected workspace for docs, wikis, and projects
Take a free 3-minute scan and get personalized AI skill recommendations.
Take free scan