Silicon

Silicon is another wrapper of Elixir/Erlang crypto packages.

Installation

The package can be installed by adding silicon to your list of dependencies in mix.exs:

def deps do
  [
    {:silicon, "~> 0.1.0"}
  ]
end

The docs can be found at https://hexdocs.pm/silicon.

Packages

We wrapped the packages below:

• {:libdecaf, "~> 1.0"} for Ed25519 and SHA3
• {:keccakf1600, "~> 2.0", hex: :keccakf1600_orig} for Keccak Origin
• {:libsecp256k1, "~> 0.1.10"} for Secp256k1
• {:blake2_elixir, git: "https://github.com/riverrun/blake2_elixir.git"} for Blake2b
• :crypto for others

Test Vectors

We added lots of extra test vectors to test against the methods exposed by silicon

AES Test Vectors

• AES-CBC: https://csrc.nist.gov/projects/cryptographic-algorithm-validation-program/block-ciphers#AES
• AES-CBC-PKCS7: https://raw.githubusercontent.com/google/wycheproof/master/testvectors/aes_cbc_pkcs5_test.json
• AES-GCM: https://csrc.nist.gov/Projects/Cryptographic-Algorithm-Validation-Program/CAVP-TESTING-BLOCK-CIPHER-MODES#GCMVS
• AES-CTR: Got from https://github.com/pyca/cryptography, which is the test vectors from RFC 3686

Hash Function Test Vectors

• SHA3: https://csrc.nist.gov/Projects/cryptographic-algorithm-validation-program/Secure-Hashing
• Keccak origin: https://keccak.team/archives.html Known-answer and Monte Carlo test results part
• Blake2b: https://github.com/jedisct1/crypto-test-vectors and https://pynacl.readthedocs.io/en/latest/vectors/blake2_vectors/
• MD5/SHA2: https://cryptii.com
• HASH160: https://bitcoinprices.org/public-key-to-hash/
• RIPEMD160: https://www.browserling.com/tools/ripemd160-hash

HMAC Test Vectors

• HMAC_SHA256/SHA512: https://cryptii.com

Ed25519 Test Vectors

• https://ed25519.cr.yp.to/python/sign.input

Secp256k1 Test Vectors

• https://github.com/btccom/secp256k1-go
• https://github.com/google/wycheproof/tree/master/testvectors

For wycheproof test vectors, We only test against ecdh vectors. For ecdsa tests, some cases might fail because that the Bitcoin-Secp256k1 seems more strict. See an example here. We don't have a good way to distinguish those cases now.