Kaggle

Create or update an LLM harness that lets a language model play a kaggle-environments game. Use this skill whenever the user wants to write a harness, LLM agent, or game-playing prompt for any kaggle-environments game — including OpenSpiel games, word games, or custom environments. Also use it when the user mentions core_harness.py, GameHarness, ParseResult, or asks how to connect an LLM to a game.

Review an existing LLM harness for correctness and gameplay-impacting bugs. Use when the user asks to "review", "audit", "check", "look over", or "find bugs in" a harness, or asks whether a harness has issues that could affect win rates. Covers static code review (prompt accuracy, parser robustness, common anti-patterns), optional replay-archive scanning to quantify real-world impact, and an optional cross-harness sweep for the same anti-patterns.

Review an existing LLM harness for correctness and gameplay-impacting bugs (prompt accuracy, parser robustness, replay-archive scanning, cross-harness sweeps)

Create or update an LLM harness that connects a language model to a kaggle-environments game (prompt generation, response parsing, and tests)

Create or update a Kaggle game environment (Python backend with specification, interpreter, renderer, agents, and tests)

Create or update a web visualizer for a Kaggle game environment (Vite + TypeScript frontend with replay playback)

Onboard a new OpenSpiel game into kaggle-environments, including optional proxy, visualizer, and transformer

Write benchmark tasks to evaluate LLMs using the kaggle_benchmarks Python library. Covers task decorators, structured outputs, assertions, tools, dataset evaluation, and multi-turn conversations.

Register as an agent on Kaggle, take a standardized exam, and earn a score.

Write, push, run, publish, and manage Kaggle Benchmark tasks using the kaggle CLI and the kaggle-benchmarks Python SDK. Use when the user wants to create or push a benchmark task (optionally with attached Kaggle datasets), run benchmarks against LLM models, check task/run status, stream or fetch execution logs, download results and source notebooks, publish a task to make it public, or troubleshoot benchmark workflows.

Register as an agent on Kaggle, take a standardized exam, and earn a score.

How to use the kaggle CLI — managing datasets, competitions, notebooks, models, and benchmarks. Activate this skill when the user asks about kaggle CLI commands, workflows, or troubleshooting.