Listing Tracked Files

When you need to see exactly which files Crab is managing in your repository — their names, sizes, content hashes, and whether they're hydrated or still pointers — crab ls-files gives you the full picture.

What It Shows

For each Crab-tracked file in your working tree, you can see:

• Content hash — The blake3 hash identifying this file version (from the pointer blob)
• Hydration state — Whether the file is a pointer (p) or hydrated (*)
• File path — Relative path in the repository
• Size — Original file size (with --size)

Basic Usage

crab ls-files

abc1234567 p models/weights.bin
def8901234 p models/embeddings.bin
---------- * data/train.csv

The p marker means the file is a pointer (dehydrated). The * marker means it's hydrated (full content on disk). The hash is ---------- for hydrated files because the pointer isn't readable when full content is present.

With File Sizes

crab ls-files --size

abc1234567 p models/weights.bin (1.2 GB)
def8901234 p models/embeddings.bin (800 MB)
---------- * data/train.csv (45.2 MB)

Just File Names

For scripting or piping to other tools:

crab ls-files --name-only

models/weights.bin
models/embeddings.bin
data/train.csv

Common Patterns

Find all pointer files (not yet hydrated)

crab ls-files | grep ' p '

Count tracked files

crab ls-files --name-only | wc -l

Feed into hydration

crab ls-files --name-only | grep 'models/' | xargs crab hydrate

JSON for automation

crab ls-files --json | jq '.data.files[] | select(.hydrated == false) | .name'

Difference from crab status

• crab status gives a summary — counts of hydrated, pointer, and modified files.
• crab ls-files gives a listing — every tracked file with its details.

Use status for a quick overview, ls-files when you need the full list.

CLI Reference

For complete command syntax, all output modes, and JSON format, see the crab ls-files reference.