NVIDIA Kaggle Plugin
The NVIDIA plugin guides an agent through a Kaggle competition: rules, winner writeups, public notebooks and solution submission.
- 350 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The plugin works with a Kaggle account through a token and can submit solutions and upload datasets.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/nvidia-kaggle-pluginDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The plugin works with a Kaggle account through a token and can submit solutions and upload datasets.
This is third-party code. Review the repository files before installing.
What it does
The plugin gives an agent one skill, `nvidia-kaggle-skill`, and picks the right script from the request. It accepts a competition slug, a Kaggle URL, a kernel ref, a writeup URL or a local path. README workflows: fetch the overview, rules, metric, timeline and dataset description; find leaderboard writeups and summarize the approaches; index discussions and notebooks and search them; download a notebook with its inputs for local reproduction; submit to a competition and poll for the result; create or update a Kaggle dataset from a local folder.
Who it is for. Kaggle participants and ML engineers who start a new task and want to quickly study what has already won.
Good fit when
- Before a new competition the overview, rules, metric and data description are collected.
- Short summaries of the top 3 solutions of a past competition are needed.
- Public notebooks are indexed and searched by technique or feature.
- A solution is submitted through a notebook and the final score is awaited.
Not a fit when
- There is no Kaggle account or `KAGGLE_API_TOKEN`.
- Only general ML reference is needed, unrelated to Kaggle.
- Submissions cannot be left to an agent without manual review.
Example request
Find the three best writeups of the AI Mathematical Olympiad Progress Prize 2 competition and briefly describe what they share.Limitations
API-backed scripts (datasets, kernels, submissions, dataset uploads) need a Kaggle account and `KAGGLE_API_TOKEN`. Python 3.10 and an agent runtime with plugin or Agent Skills support are required. Submitting to a competition and uploading a dataset change your Kaggle account, and daily submissions are limited by competition rules. Accepting competition rules on Kaggle remains the user's job, and submission fails without it. Writeups and notebooks are summarized by a model, so techniques and numbers are worth checking against the source. The last commit is from August 2026.
How to disable. Uninstall the `nvidia-kaggle` plugin in the agent's plugin manager and remove `KAGGLE_API_TOKEN` from the environment. Submitted solutions and datasets are deleted on the Kaggle site.
FAQ
Which skill is inside the plugin?
One, `nvidia-kaggle-skill`. It selects a script from the request, whether a slug, a URL or a local path.
How do I call the skill in Claude Code?
With a command like `/nvidia-kaggle:nvidia-kaggle-skill` followed by the task, for example summarizing the best solutions of a competition.
Is a token required?
`KAGGLE_API_TOKEN` is needed for API workflows: datasets, kernels, submissions, uploads. Some read-only workflows run without it.
Which agents are supported?
The README covers Codex and Claude Code, both through a plugin marketplace.
What does solution submission do?
The script pushes a kernel to the competition and polls for the final result. Before sending, the agent should show what exactly goes to Kaggle.
Discussion
No questions or reviews yet.
Related
Pluginskaggle-skill answers Kaggle competition questions with sources, tracks your standing and submits only after you agree.
A skill that runs a Kaggle competition from rules and folds to a scored submission, with hidden-rerun debugging and ensembles
For a repeatable experiment, an agent downloads a Kaggle dataset to a working directory and retrieves notebook outputs.
NVIDIA BioNeMo Agent Toolkit: skills for life sciences
NVIDIA BioNeMo Agent Toolkit
NVIDIA skills for biology and chemistry: the agent predicts protein structures, docks molecules and designs proteins
A skill set for math modeling contests: problem parsing, method selection, Python and MATLAB code, and paper writing
An official Google Cloud plugin: the agent finds BigQuery tables, writes SQL and builds forecasts from natural language questions.
README in short
The README lists workflows in a table, shows examples with answers, requirements, marketplace install for Codex and Claude Code, and a local install.