Storyblok CLI for coding agents

If a coding agent needs to inspect media stored in a Storyblok space, it can use the storyblok command-line interface (CLI) to download assets and their metadata into a local directory. Its documented assets pull command also has a --dry-run option, which previews changes without applying them locally. The agent can review the preview before it writes any files.

The CLI lets an agent that runs commands in a shell access Storyblok. The downloaded files give the agent copies it can inspect locally. The example below previews the pull first and downloads files in a separate step. See Storyblok’s CLI documentation for current installation and authentication instructions.

Preview an asset pull

Storyblok documents assets pull for downloading assets, asset metadata, and asset folders from a space. The command requires a numeric space ID. Replace YOUR_SPACE_ID below with the ID for the space you intend to inspect:

SPACE_ID=YOUR_SPACE_ID
storyblok assets pull --space "$SPACE_ID" --dry-run

The CLI asks Storyblok which assets the pull would select. The --dry-run flag keeps the command from changing the local directory. Read the CLI’s preview output before running the command without --dry-run. The asset pull reference explains the flag but does not promise a fixed output format. Ask the agent to report what the command prints instead of expecting a particular format.

The command reference also lists --target values for choosing which assets to pull: with-referenced, all, space, or shared. The default is with-referenced. Choose another target only when the task needs a different set of assets. For example, an agent asked to review assets used by a specific content change should start with the default rather than download every asset in the space.

Download files for local inspection

After reviewing the preview, run the pull without --dry-run to write files locally:

storyblok assets pull --space "$SPACE_ID"

This step changes the local directory. Storyblok stores space assets under a folder named for the space ID and shared-library assets under shared/<library-id>/. Use --path to choose where the CLI writes the files. Check that destination before the pull, especially if it already contains files: running without --dry-run applies changes locally and may overwrite existing files.

The CLI also supports --query to filter the pull. Use a query only when you know it fits the task. The command documentation lists the option but does not give one query value that fits every task. For private assets, the reference documents --asset-token. Treat that token as a secret and provide it through the agent’s protected runtime configuration, not in its prompt or checked-in command history.

Configure access before the agent runs

The asset command requires a space ID, and private assets require the documented asset token. The assets pull reference does not describe general CLI sign-in, credential storage, or the installation command. Follow the current Storyblok CLI documentation for those setup details instead of guessing a login command or embedding credentials in an agent instruction.

You should install and configure the CLI for the intended Storyblok space before asking the coding agent to run the preview. If access fails, the operator should check the CLI’s documented authentication setup and confirm that the configured credentials cover the requested space and assets.

Give the agent a bounded task

A useful prompt names the space, tells the agent to preview the pull, and asks it to stop before downloading files:

Preview an asset pull for Storyblok space YOUR_SPACE_ID with storyblok assets pull --space "$SPACE_ID" --dry-run. Summarize the preview output and stop. Do not download files or include credentials in your response.

Once a person approves the preview, the agent can run the command without --dry-run and inspect the resulting local files. The CLI documentation covers this asset workflow; it does not establish here how other Storyblok content types or migrations should be handled.