What problem does it solve? Publishing large binary dataset artifacts (such as LanceDB snapshots) to Hugging Face Git repositories often fails with binary rejection errors, inconsistent tracking rules, and diverged branches. This Skill provides a complete operational playbook for setting up Git Xet tracking, committing safe snapshots, and recovering from failed pushes. ## Core Features & Use Cases - Repository Setup & Bootstrap: Initialize or bind a local data folder to a Hugging Face dataset remote, whether the remote already has history or is empty, and configure Git Xet tracking for binary extensions like .lance, .txn, .manifest, and .idx. - Daily Snapshot Workflow: Commit and push dataset updates safely with preflight checks for branch divergence, remote state, and tracking rules. - Binary Rejection Recovery: Diagnose and fix pre-receive hook rejections caused by untracked binary files, including branch surgery and clean-branch publication strategies. - Use Case: You maintain a LanceDB database synced to a Hugging Face dataset repo. A push fails with "Your push was rejected because it contains binary files" for .idx files. Use this Skill to add the missing Xet track rule, realign local main to the remote tip, re-commit, and push successfully. ## Quick Start Set up my local LanceDB data folder as a Hugging Face dataset Git repository with Xet tracking and push the initial snapshot to the remote.
