From b8516b7fa6c423d164ffc62433b3d8036e8136ff Mon Sep 17 00:00:00 2001 From: nicholaskang-us Date: Fri, 15 May 2026 09:05:31 -0700 Subject: [PATCH] Update and rename benchmarks.md to SKILL.md --- skills/references/{benchmarks.md => SKILL.md} | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) rename skills/references/{benchmarks.md => SKILL.md} (94%) diff --git a/skills/references/benchmarks.md b/skills/references/SKILL.md similarity index 94% rename from skills/references/benchmarks.md rename to skills/references/SKILL.md index 54c6d595..3eca6605 100644 --- a/skills/references/benchmarks.md +++ b/skills/references/SKILL.md @@ -1,7 +1,26 @@ +--- +name: kaggle-benchmarks +description: > + How to write, push, run, and manage Kaggle Benchmark tasks using the kaggle + CLI and the kaggle-benchmarks Python SDK. Activate this skill when the user + wants to create a benchmark task, push a task file, run benchmarks against + LLM models, check run status, download results, or troubleshoot benchmark + workflows. Keywords: kaggle benchmarks, benchmark task, kbench, model proxy, + push task, run task, benchmark status, benchmark download. +metadata: + author: kaggle + version: "0.1" +--- + # Kaggle Benchmarks CLI Reference This reference covers how to use the `kaggle` CLI to manage Kaggle Benchmark tasks — pushing task files, running them against LLM models, checking status, and downloading results. +## Official resources + +- **kaggle-benchmarks SDK repo:** https://github.com/Kaggle/kaggle-benchmarks — full source, API reference, and examples for the `kaggle-benchmarks` Python library used to write task files +- **DeepWiki documentation:** https://deepwiki.com/Kaggle/kaggle-benchmarks — auto-generated documentation for the SDK + ## Prerequisites - Python 3.11+