kaggle-workflow
Guide through complete Kaggle competition workflow from TODO updates to submission. Use PROACTIVELY when user starts new experiments, prepares submissions, or asks about competition workflow. Keywords: Kaggle, 提出, submission, workflow, ワークフロー, コンペ
When & Why to Use This Skill
This Claude skill streamlines the end-to-end Kaggle competition lifecycle, from initial experiment setup to final submission. It helps data scientists maintain rigorous documentation, implement robust cross-validation strategies, and manage the transition from training pipelines to submission-ready notebooks, ensuring a structured and reproducible machine learning workflow.
Use Cases
- New Competition Kickoff: Automatically setting up a structured TODO.md and project roadmap when starting a new Kaggle challenge.
- Experiment Tracking: Managing the pipeline from model training to evaluation using best practices for documentation and result logging.
- Submission Preparation: Generating and checking submission notebooks to ensure they meet competition requirements and avoid common errors.
- Validation Strategy Design: Implementing and documenting cross-validation patterns to ensure model reliability and prevent leaderboard overfitting.
| name | kaggle-workflow |
|---|---|
| description | Guide through complete Kaggle competition workflow from TODO updates to submission. Use PROACTIVELY when user starts new experiments, prepares submissions, or asks about competition workflow. Keywords: Kaggle, 提出, submission, workflow, ワークフロー, コンペ |
When to Use (PROACTIVE)
This skill should be activated automatically when:
- User starts a new Kaggle competition
- User prepares for submission
- User asks about competition workflow
- User mentions: "Kaggle", "提出", "submission", "ワークフロー"
- User needs guidance on experiment → submission pipeline
What This Skill Does
Provides end-to-end Kaggle competition workflow:
- TODO.md management best practices
- Training → Trackio → Evaluation pipeline
- Submission notebook preparation
- Cross-validation strategies
- Experiment documentation patterns
How to Use
Refer to submission-guide.md for:
- Complete workflow from start to submission
- TODO.md integration with experiments
- Submission notebook checklist
- Cross-validation best practices
- Experiment documentation templates