casey/filepack

filepack

Filepack is a content-addressed package format with self-describing metadata.

A package is a directory of files and a manifest containing their BLAKE3 hashes. File and directory hashes form a Merkle tree whose root hash, the package fingerprint, uniquely identifies the package. Packages can be signed with Ed25519 and verified against a manifest, fingerprint, or signature.

Packages may contain machine-readable metadata describing their content, allowing programmatic search, preview, playback, and conversion.

filepack is a command-line tool for creating, signing, and verifying packages, and includes an HTTP server for package upload, download, and display.

A demo server with a selection of example packages is available at filepack.com.

Filepack is currently experimental. The filepack interface and package format may change at any time.

Quickstart

Optionally, add metadata describing your package:

echo "title: Packaging Guide" > metadata.yaml

Create manifest.filepack:

filepack create

Verify current directory against hashes in manifest.filepack:

filepack verify

Print package fingerprint:

filepack fingerprint

Print package metadata:

filepack metadata

Installation

filepack is written in Rust and can be built from source and installed from a checked-out copy of this repo with:

cargo install --path .

Or from crates.io with:

cargo install filepack

See rustup.rs for installation instructions for Rust.

Pre-Built Binaries

Pre-built binaries for Linux, MacOS, and Windows can be found on the releases page.

You can use the following command on Linux, MacOS, or Windows to download the latest release, just replace DEST with the directory where you'd like to put filepack:

curl --proto '=https' --tlsv1.2 -sSf https://filepack.com/install.sh | bash -s -- --to DEST

For example, to install filepack to ~/bin:

# create ~/bin
mkdir -p ~/bin

# download and extract filepack to ~/bin/filepack
curl --proto '=https' --tlsv1.2 -sSf https://filepack.com/install.sh | bash -s -- --to ~/bin

# add `~/bin` to the paths that your shell searches for executables
# this line should be added to your shell's initialization file,
# e.g. `~/.bashrc` or `~/.zshrc`
export PATH="$PATH:$HOME/bin"

# filepack should now be executable
filepack --help

Note that install.sh may fail on GitHub Actions or in other environments where many machines share IP addresses. install.sh calls GitHub APIs in order to determine the latest version of filepack to install, and those API calls are rate-limited on a per-IP basis. To make install.sh more reliable in such circumstances, pass a specific tag to install with --tag.

Usage

Filepack supports a number of subcommands, including filepack create to create a manifest, filepack verify to verify a manifest, and filepack serve to start an HTTP package server.

See filepack help for supported subcommands and filepack help SUBCOMMAND for information about a particular subcommand.

filepack create

Create a manifest.

Recommended lints can be enabled with:

filepack create --deny distribution

filepack verify

Verify the contents of a directory against a manifest.

To verify the contents of DIR against DIR/manifest.filepack:

filepack verify DIR

If the current directory contains manifest.filepack, DIR can be omitted:

filepack verify

filepack verify takes an optional --print flag, which prints the manifest to standard output if verification succeeds. This can be used in a pipeline to ensure that you the manifest has been verified before proceeding:

filepack verify --print | jq

filepack serve

Start HTTP server. The filepack server has limited authentication and no disk usage quota and should be considered experimental.

To serve HTTP on 0.0.0.0:80:

filepack serve

To serve HTTPS on 0.0.0.0:443 with an ACME TLS certificate:

filepack serve --https --domain filepack.example

See filepack serve --help for more details.

Files can be uploaded with filepack upload and downloaded with filepack download.

Data Directory

Filepack stores local configuration in the filepack subdirectory of the platform data directory.

The location of the platform data directory is platform dependent, with $XDG_DATA_HOME being respected on all platforms.

Filepack stores keys in the keychain subdirectory of the filepack data directory. So if the platform data directory is ~/.local/share, the keychain directory is ~/.local/share/filepack/keychain.

The location of the filepack data directory used by a command can be overridden with the --data-dir option.

Manifest

filepack manifests are conventionally named manifest.filepack and are placed alongside the files they reference.

Manifests are CBOR and may be converted to JSON for inspection or manipulation with filepack manifest.

Manifests, when converted to JSON, are an object with three mandatory keys, embedded, package, and signatures.

embedded

The value of the mandatory embedded key is an object mapping BLAKE3 hashes to hex-encoded file contents. The package object only includes file names, hashes, and sizes, but files may be embedded in the manifest archive in the embedded map.

This is currently only used to embed metadata.filemeta.

package

The value of the mandatory package key is an object mapping path components to directory entries. Directory entries may be subdirectories or files. Files are objects with keys hash, the hex-encoded BLAKE3 hash of the file, and size, the length of the file in bytes.

Path components are UTF-8 and may not be . or .., contain the path separators / or \, contain control characters, be longer than 255 bytes, or begin with a Windows drive prefix, such as C:.

signatures

The value of the mandatory signatures key is an array of signatures. Signatures are Bech32m strings that include an Ed25519 key, the package fingerprint the signature is made over, an optional timestamp, and the signature itself.

Public keys are Curve25519 points and signatures are Ed25519 signatures made over the hash of a serialized CBOR statement containing the package fingerprint which commits to the content of package.

Example

An manifest converted to JSON over a directory containing the files README.md and src/main.c, signed by the public key public1a67dndhhmae7p6fsfnj0z37zf78cde6mwqgtms0y87h8ldlvvflyqcxnd63:

{
  "embedded": {},
  "package": {
    "README.md": {
      "hash": "fc253b84551ce6b00e820a826ac18054dc7f63a318ce62f3175315f5c467a62a",
      "size": 11883
    },
    "src": {
      "main.rs": {
        "hash": "1fa48b95ed335369d45b91af8138bdccd1413364bcdbfa6e9034e8a2cfd6e17f",
        "size": 33
      }
    }
  },
  "signatures": ["…"]
}

The signature is elided for brevity. Signatures are Bech32m-encoded strings containing both a public key and an Ed25519 signature.

Keys, Signatures, Fingerprints, and Hashes

Public keys, private keys, signatures, and package fingerprints are all Bech32m-encoded strings beginning with public1…, private1…, signature1…, and package1… respectively.

BLAKE3 file hashes are 64-character lowercase hexadecimal.

Metadata

Filepack packages may contain a file named metadata.filemeta describing the package and its contents. If present, metadata.filemeta is also embedded in the manifest.filepack archive.

Metadata is authored by creating a file named metadata.yaml in the root of a new package. filepack create then loads metadata.yaml if present, checks for validity and unknown fields, and writes the CBOR serialization to metadata.filemeta in the package root.

metadata.yaml is retained as a human-readable reference and for amending metadata, but metadata.filemeta is the authoritative source of metadata. For consumption by scripts and tools, filepack metadata prints the contents of metadata.filemeta as JSON.

Filepack metadata is intended to be a broadly useful machine and human readable description of the contents of a package, covering personal, distribution, and archival use-cases.

Metadata follows a fixed schema and is not user-extensible. Future versions of filepack may define new metadata fields, causing verification errors if those fields are present and invalid according to the new schema.

Please feel free to open an issue with ideas for new metadata fields.

Schema

This schema is for the YAML authoring format and the JSON output of filepack metadata. The CBOR schema is currently undocumented.

Fields are given as NAME: TYPE. All fields are optional.

Top-level fields

• artwork: component.{jpeg,png}: The filename of a JPEG or PNG file containing artwork for the content, for example, cover art for an album or key art for a movie.

• creator: component: The person or group who created the content.

• description: markdown: A description of the content.

• homepage: url: Primary URL for the content. Should be the official homepage of the content, if any, and not, for example, a Wikipedia or media database link.

• language: language: The primary language of the content.

• package: object: The package metadata.

• readme: component.md: The filename of the content readme.

• time: time: The time the content was created or released.

• title: component: The content's human-readable title.

Fields of package describing the package itself, as opposed its content:

• colophon: component.md: The filename of the package colophon.

• creator: component: The person or group who created the package.

• description: markdown: A description of the package.

• homepage: url: Primary URL for the package.

• time: time: The time the package was created.

• title: component: The package's title. May contain details related to the source, packager, or encoding. May to be used as a directory name when saving the package if the main title field would be ambiguous.

Types:

• component: A string with the same restrictions as path components in the manifest package object, allowing them to be used as unix filesystem paths. Note that Windows imposes additional restrictions which are not enforced, so components may not be valid paths on Windows.

• component.EXTENSION: A component that must end with .EXTENSION.

• language: A string containing an ISO 639-1 two-character language code. See filepack languages for valid language codes.

• markdown: A string containing CommonMark markdown.

• time: A string containing a time with one of two precisions, a year only, or a year, month, and day, written Y-MM-DD. Times use the proleptic Gregorian calendar with astronomical year numbering.

• url: A string containing a URL with scheme http or https.

Example times:

0
-44
1929
-13787000000
0-01-01
929-01-01
1970-01-01
-44-03-15

Example

title: Tobin's Spirit Guide
creator: John Horace Tobin
artwork: cover.png
time: 1929
description: A compilation of supernatural occurrences, entities, and facts.
homepage: https://tobin-society.org/spirit-guide
language: en
readme: README.md
package:
  title: Tobin's Spirit Guide - First Edition
  creator: Egon Spengler
  time: 1984-07-08
  description: >
    First edition on loan from NYPL Main Branch research stacks. Captured via
    Microtek MS-300A flatbed scanner.
  homepage: https://ghost-busters.net/~egon
  colophon: COLOPHON.md

The homepage URLs are of course anachronistic, as the World Wide Web was created in 1989, some years after Egon first packaged Tobin's Spirit Guide.

Lints

filepack create supports optional lints that can be enabled by group:

filepack create --deny distribution

The distribution lint group checks for issues which can cause problems if the package is intended for distribution, such as non-portable paths that are illegal on Windows, paths which would conflict on case-insensitive file systems, and inclusion of junk files such as .DS_Store.

Lint group names and the lints they cover can be printed with:

filepack lints

Keys and Signatures

filepack supports the generation of Curve25519 public/private keypairs, and the creation and verification of EdDSA signatures over manifests.

Keypair Generation

Keypairs are generated with:

filepack keygen

Which creates master.public and master.private files in the keychain subdirectory of the filepack data directory.

Public Key Printing

Generated public keys can be printed with:

filepack key

Signing

Signatures are created with:

filepack sign

Which signs the manifest in the current directory with your master key and adds the signature to the manifest's signatures array. Signatures are made over a fingerprint hash, recursively calculated from the contents of the manifest.

Signature Verification

Signatures embedded in a manifest are verified whenever a manifest is verified. The presence of a signature by a particular public key can be asserted with:

filepack verify --key PUBLIC_KEY

Which will fail if a valid signature for PUBLIC_KEY over the manifest contents is not present.

Fingerprints

Filepack signatures are made over the package fingerprint, which is the hash of a CBOR object which commits to the files and directories contained in the manifest.

Fingerprints are BLAKE3 hashes, constructed such that it is impossible to produce packages which are different but which have the same fingerprint.

Fingerprints may be used as a globally unique identifier. If two packages have the same fingerprint, they have the same content.

Guide

The [DIRECTORY] argument in examples can be omitted, in which case it defaults to the current directory.

Creating a Package

To create a filepack manifest:

filepack create [DIRECTORY]

This creates manifest.filepack containing hashes and file sizes of all files in the DIRECTORY and subdirectories.

To enable linting, use the --deny flag with a lint or lint group:

filepack create --deny <LINT> [DIRECTORY]

To view all lints:

filepack lints

Please feel free to open issues requesting new lints.

Creating a Package with Metadata

To create a package with metadata describing the package, create a file named metadata.yaml in the package directory, for example:

title: The Necronomicon
creator: Abdul Alhazred
description: >
  The Old Ones, their history, and the rites by which they may be summoned.
language: la

Then, to create the package:

filepack create [DIRECTORY]

Which will validate metadata if present and write it to metadata.filemeta, as well as include it in manifest.filepack.

See metadata for the full schema.

Signing a Package

To sign a package when creating it, first create a new public and private key pair:

filepack keygen

This creates a new public key, master.public, with corresponding private key, master.private, in the filepack keychain directory.

The filepack info command will print the path to the filepack data directory, keychain directory, and any public keys in the keychain directory.

Your public key can be printed with:

filepack key

To sign a new package:

filepack create --sign [DIRECTORY]

To add a signature to an existing package:

filepack sign

Getting the Package Fingerprint

To print the fingerprint of a package:

filepack fingerprint [DIRECTORY]

Publishing a Manifest

Manifests contain only hashes and not file content, and so can be published anywhere, including places where technical or legal limitations would preclude publication of the content itself.

Verifying a Package

Package content can be verified with:

filepack verify [DIRECTORY]

Any extra files, missing files, or modified files will produce an error.

This verifies files against hashes stored in the manifest.

Running filepack verify on its own can detect accidental corruption, but not intentional modification, since an attacker could modify package content, and then modify the manifest to make it match the modified content and thus pass verification.

However, if the manifest was obtained from a trusted source, verification will catch any modifications, intentional or otherwise, even if the package content was obtained from an untrusted source.

To detect intentional modification, you can either verify a package against a known-good fingerprint:

filepack verify --fingerprint <FINGERPRINT> [DIRECTORY]

This verifies both the contents of the package and that the manifest has the expected fingerprint. The former protects against accidental corruption, and the latter against intentional modification.

Packages can also be verified by checking for a signature by a known public key. This also protects against intentional modification, as long as the private key corresponding to the public key is protected, and only known to you, or someone you trust.

To verify that a package has a signature from a particular public key:

filepack verify --key <PUBLIC_KEY> [DIRECTORY]

The --key option can be repeated to require signatures from multiple public keys.

Detecting Accidental Corruption

Create a filepack manifest with:

filepack create <PACKAGE>

This will create <PACKAGE>/manifest.filepack

To later verify the package against the manifest:

filepack verify <PACKAGE>

Because the manifest contains cryptographic hashes, accidental corruption to the files or manifest will always be detected by filepack verify.

This is not the case with intentional, malicious corruption, since an attacker could modify the files and replace the manifest hashes with the hashes of the modified files.

Detecting Malicious Corruption

Because an attacker could modify the files and replace the manifest hashes with the hashes of the modified files, you must ensure that the manifest has not been tampered with.

This can be accomplished in a number of ways, either by saving the manifest to a secure location, saving the package fingerprint, or signing the package.

Manifest

To save the manifest in a secure location, use the --manifest option to save the manifest somewhere other than the package:

filepack create <PACKAGE> --manifest <MANIFEST>

Then, verify the package against the saved manifest:

filepack verify <PACKAGE> --manifest <MANIFEST>

Because the manifest was protected, any modification to the package will be detected. This has the advantage that not only will any modifications be detected, but which files were modified can also be detected.

Fingerprint

Create the manifest in the package root with:

filepack create <PACKAGE>

Print the package fingerprint:

filepack fingerprint <PACKAGE>

Save the fingerprint in a secure location.

Then, verify the package against the saved fingerprint:

filepack verify <PACKAGE> --fingerprint <FINGERPRINT>

Because the fingerprint was protected, any modification to the package will be detected. This has the advantage that you only have to save a small text string, but the disadvantage that while any modifications will be detected, you will not be able to determine which files have changed.

Signature

Create the manifest in the package root and sign it with your master key:

filepack create <PACKAGE> --sign

Then, verify the package and its signature:

filepack verify <PACKAGE> --key master

Any modification to the package or manifest will invalidate the signature, which will be detected. This has the advantage of not needing to save the manifest or fingerprint of packages you want to verify. However, you will need to generate and secure your private key.

Determining Authenticity

To check the authenticity of a package created by someone else, get their public key and verify that the package contains a signature by that key:

filepack verify <PACKAGE> --key <KEY>

Alternatives and Prior Art

filepack serves the same purpose as programs like shasum, which hash files and output a text file containing file hashes and paths, which can later be used with the same program to verify that the files have not changed.

They output hashes and paths one per line, separated by whitespace, and mainly differ in which hash function they use.

Some examples, with links to implementations and the hash functions they use:

CRC-32 is not a cryptographic hash function and cannot be used to detect intentional modifications. Similarly, SHA-1 was thought to be a cryptographic hash function, but is now known to be insecure.

filepack and b3sum both use BLAKE3, a fast, general-purpose cryptographic hash function.

filepack can also create and verify signatures. Other signing and verification utilities include:

Filepack for SFV Enjoyers

If you package content for distribution, Filepack offers a number of benefits over simple file verification with .sfv files.

• Filepack can detect both accidental corruption and intentional modification, whereas .sfv files can only detect accidental corruption.

• Because manifest.filepack manifests contain file sizes, Filepack can tell the user not just whether a file has been modified, but also whether it is empty, truncated, or too long.

• Filepack packages have a fingerprint, a short text string beginning with package1… which is guaranteed to be globally unique, allowing packages to be identified and referenced by fingerprint alone.

• Fingerprints can be used to verify that a manifest.filepack manifest itself has not been tampered with, proving authenticity of a package regardless of its source.

• Packages can be signed, allowing users to verify authenticity of any package from a packager by public key, a short string beginning with public1….

• Filepack can warn you if package filenames might cause issues on other operating systems or file systems.

• Filepack packages are Merkle trees, both of files and within files. A user with a manifest, package fingerprint, or trusted public key can incrementally stream and verify files, or access and verify file content at random. Additionally, errors can be detected and recovered from by transmitting only those parts of the file which are corrupted.

• Packages may contain machine readable metadata following a filepack-defined schema, allowing packages to be searched, indexed, and exposed through rich, featureful interfaces.

Why is metadata important?

Metadata is the difference between a good user experience and a bad user experience.

On Netflix, movies are presented in an attractive fashion with artwork, titles, actors, and other information. All content is playable on whatever device you happen to be using, streaming starts instantly, search is useful, and a recommendation engine surfaces movies you might like. It is usable by anyone, of any age and with any degree of technical proficiency.

On BitTorrent, movies are presented inscrutably in text listings with weirdly formatted titles, spotty or missing information, and no artwork. Content may or may not be playable on your device, streaming is impossible, search is unstructured, and recommendations are unavailable. It is only usable by a small and decreasing percentage of the population, and asking someone over to "BitTorrent and chill" will be met by blinking incomprehension, followed by derision as you fiddle with your torrent client like a digital steam engine operator, desperately trying to find seeds for the movie you want to watch.

The difference between the two systems is not due to a difference in underlying data, every piece of content one could possibly imagine is available on BitTorrent. It is a difference in the availability of metadata.

Netflix has a database of standardized, machine-readable metadata, supporting nearly every feature of the system. BitTorrent has a disparate collection of folders of files, with inconsistent and incomplete metadata, and the user experience and features of the system are a direct consequence.

As polished services driven by metadata have become more popular, the user experience that systems based on folders of files can offer has become ever more unacceptable and alien, regardless of cost or underlying availability of content.

Filepack seeks to rectify this by standardizing machine-readable metadata that can be included with folder-of-files content.

Filepack metadata is stored a file, metadata.filemeta, in the root of a filepack package as well as in the manifest.filepack archive, and contains information related to what the package contains, which files in the package contain the package content and their file formats, and who created the package. This metadata can serve as a base for the creation of rich local and distributed applications and services, with user experiences that compete with and exceed those of closed centralized alternatives.

Design

Manifest Format

A filepack manifest contains all information needed to verify the contents of a directory. The package key of the manifest is a directory object mapping filenames to directory entries, which may themselves be directories, or files, in which case they contain the hash of the file contents, as well as the length of the file.

The length of the file is not strictly necessary for verification, but is included so that truncated, empty, and overlong files can be identified, which may help in understanding verification failures.

File contents may be embedded in the manifest.filepack archive under the embedded key.

File Hashes

The contents of files are hashed with BLAKE3 using the official Rust implementation. BLAKE3 was chosen both for its speed, and for the fact that it utilizes a Merkle tree construction. A Merkle tree allows for verified file streaming and subrange inclusion proofs, which both seem useful in the context of file hashing and verification.

Signatures

Filepack allows for the creation of Ed25519 signatures over the contents of a manifest, which thus commit to the contents of the directory covered by the manifest. Signatures are made over a statement containing a "fingerprint" hash of a canonical CBOR serialization of the manifest. This keeps signatures independent of the manifest format, avoids issues with canonicalization of the manifest JSON, avoids hash loops due to the inclusion of signatures in the manifest itself, and allows proving the inclusion of files covered by a signature.

Fingerprints

Package fingerprints are the BLAKE3 hash of a canonical CBOR serialization of the contents of the manifest. Fingerprints are constructed to be unique, meaning that it is impossible for two different packages with different contents to have the same fingerprint.

Currently, no fingerprint test vectors exist, and the best documentation is the code itself.

In particular, see:

• Manifest
• Directory
• Entry
• Files
• Statement

And the cbor module for the encoding.