This walkthrough takes a directory of generated output, turns it into aDocumentation Index
Fetch the complete documentation index at: https://filepacks.com/docs/llms.txt
Use this file to discover all available pages before exploring further.
.fpk artifact, and shows how to review it.
The same pattern works for agent runs, eval snapshots, CI reports, and any other workflow that emits files.
Install the CLI
Quick trial from a clean shell:PATH, install it globally:
npx filepacks ... so the walkthrough stays copy-pasteable from a clean local state.
Create example output directories
Pack each directory
pack writes a deterministic .fpk archive and prints a summary like:
pack creates one portable .fpk file per directory. The archive digest is printed as sha256:<hex>.
Inspect what you created
inspect when you want a quick summary: artifact name, format version, digest, file count, and total bytes.
Verify integrity before trusting it
verify before sharing an artifact, uploading it, or using it as a baseline. It confirms that the payload still matches the manifest.
Compare baseline and candidate
compare exits with status 20 and prints a structural summary:
compare exits 0 with ok=true.
In automation, the exit code matters:
0means the candidate matches the baseline20means one or more packaged files changed1means the command was used incorrectly or an artifact could not be read
What to look at next
Whencompare reports a change:
- Use the
changed_file=,added_file=, andremoved_file=lines to find the affected paths. - Review the original output directories or the files inside the
.fpkartifact in your own tooling. - If the change is acceptable, store the candidate artifact explicitly as the next baseline. The public OSS CLI does not manage baselines for you.
What you just used
filepacks pack <input> --output <file>filepacks inspect <file>filepacks verify <file>filepacks compare <baseline> <candidate>
Next steps
- Read Why filepacks for when the artifact model is useful.
- Use CLI workflows for repeatable patterns in scripts, CI, and short AI-assisted-work demos.
- Read Use cases for eval, agent, and review scenarios.
- Use Artifacts when you need format-level details.