SKILL.md
Link Project
When the user mentions a project repository they've built:
Instructions
- Read the current PROJECTS.md file
- Determine the appropriate category:
- Machine Learning - Classic ML algorithms, supervised/unsupervised learning - Deep Learning - Neural networks, computer vision, NLP models - AI Engineering - LLM applications, agents, RAG systems - ML Systems & Production - MLOps, deployment, monitoring, data pipelines - Software Engineering - General software projects, tools, libraries - Other - Anything that doesn't fit above
- Add the project link in format:
``markdown - [Project Name](repo-url) - Brief description of what it does/demonstrates ``
- Maintain organization:
- Group by category (create category header if it doesn't exist) - Within each category, newest projects at the bottom - Keep descriptions concise (one line)
Examples
## Deep Learning
- [CNN Image Classifier](https://github.com/user/cnn-classifier) - From-scratch CNN implementation for CIFAR-10
- [Transformer from Scratch](https://github.com/user/transformer) - Educational implementation of attention mechanism
## AI Engineering
- [RAG Document QA](https://github.com/user/rag-qa) - Retrieval-augmented generation system with vector database
Edge Cases
- If PROJECTS.md is empty: Create basic structure with category headers
- If project fits multiple categories: Use primary category or ask user
- If repo URL not provided: Ask user for the link
- If description not clear: Ask user what the project demonstrates