Blobsy

Blobsy Agent Skill

Git-native large file storage CLI. Track large files with .bref pointers in Git while storing blobs in S3, Google Cloud Storage, Azure (via rclone), local directories, or custom backends.

Installation

If blobsy is not installed, run:

npm install -g blobsy@latest
blobsy setup --auto s3://bucket/prefix/

When to Use

Use blobsy when:

• A repository contains large binary files (models, datasets, media, archives)
• Files need to be shared across machines without committing them to Git
• You need content-addressable, deduplicated blob storage
• The user mentions: blobsy, large files, blob storage, .bref, bref, externalize, Git LFS alternative

Quick Reference

# Recommended workflow
blobsy add <path...>       # Track + stage to git (recommended)
blobsy push [path...]      # Upload blobs to backend
blobsy pull [path...]      # Download blobs from backend

# Low-level tracking (no git staging)
blobsy track <path...>     # Track files (creates .bref pointers)

# Diagnostics
blobsy status --json       # Current state (JSON)
blobsy doctor --json       # Health check (JSON)

# Configuration
blobsy config [key] [val]  # Get/set config (supports --global, --show-origin, --unset)

# Documentation
blobsy docs --readme       # Display README
blobsy docs [topic]        # User guide (--list for topics, --brief for summary)

Key Options

• blobsy add --min-size <size>: Override externalization threshold (e.g. 100kb, 5mb)
• blobsy doctor --fix: Auto-fix detected issues
• blobsy rm --remote --force: DANGER: delete the blob from the backend (breaks git history)

Current State

For dynamic info, use:

• blobsy status --json - Tracked files, sync state
• blobsy doctor --json - Configuration, health, issues

All commands accept: --json, --quiet, --verbose, --dry-run