vibgrate/cli repository overview

⁠Vibgrate CLI on Docker

Run Vibgrate against a repository without installing Node.js on your host. The CLI builds a deterministic code graph, measures dependency and runtime drift, and helps you inspect change impact and upgrade priorities. The repository stays mounted on your machine; generated .vibgrate/ files remain in that repository unless you choose --no-local-artifacts for a scan.

⁠Quick start

From the root of the repository you want to analyse:

docker run --rm \
  --mount type=bind,source="$(pwd)",target=/workspace \
  --workdir /workspace \
  --entrypoint vg \
  vibgrate/cli:latest scan

This prints your DriftScore and ranked upgrade priorities. The bind mount lets the CLI read the repository and write its local scan artifacts. Add .vibgrate/ to your .gitignore if you don't intend to commit them.

Build and query the local code map using the same mount:

docker run --rm --mount type=bind,source="$(pwd)",target=/workspace \
  --workdir /workspace --entrypoint vg vibgrate/cli:latest build

docker run --rm --mount type=bind,source="$(pwd)",target=/workspace \
  --workdir /workspace --entrypoint vg vibgrate/cli:latest \
  ask "what are the main entry points?"

For repeatable CI runs, replace latest with a published version tag such as 2026.915.1. The explicit --entrypoint vg makes the examples independent of the image's configured default command; it requires the published image to include vg on its PATH.

⁠What you can run

Use the same Docker command pattern, replacing the final scan with the command and arguments you need. Some features use network data or separate optional modules; scan --offline supports a supplied local version bundle. Cloud upload occurs only when you explicitly use a push command or flag.

For assistant integration, local models, hosted model setup, evidence workflows, flags and limitations, see the full CLI README⁠ and CLI documentation⁠. The VS Code extension⁠ provides the editor interface.

Before publishing these examples: confirm that the pushed Docker image exposes the vg executable on PATH and can write to /workspace when mounted. The supplied GitHub README does not document the image configuration.