Feat/rc6

[thomasxm]

Summary (updated)

Adds RC6 block cipher implementation to CyberChef:

• RC6 - AES finalist by Ron Rivest, Matt Robshaw, Ray Sidney, and Yiqun Lisa Yin

New Operations

RC6 (AES Finalist)

• RC6 Encrypt - Configurable block encryption (RC6-w/r/b)
• RC6 Decrypt - Configurable block decryption (RC6-w/r/b)
• Word sizes: 8, 16, 32, 64, 128, or any multiple of 8 (8-256 bits)
• Rounds: 1-255 (default 20 for standard RC6-32)
• Key length: Any (common: 128/192/256-bit)
• Modes: ECB, CBC, CFB, OFB, CTR

Padding Options

• PKCS5 (PKCS#7)
• NO (none - requires block-aligned input)
• ZERO
• RANDOM
• BIT (ISO/IEC 7816-4)

Use Cases

• General-purpose encryption
• Legacy system compatibility
• AES alternative (RC6-32/20 uses same block/key sizes)
• Educational purposes (studying AES competition finalists)

Files Changed

File	Action
src/core/lib/RC6.mjs	NEW
src/core/operations/RC6Encrypt.mjs	NEW
src/core/operations/RC6Decrypt.mjs	NEW
src/core/config/Categories.json	Added operations
tests/operations/tests/RC6.mjs	NEW
tests/operations/index.mjs	Registered tests

Testing

• All IETF draft test vectors pass (RC6-8, RC6-16, RC6-32, RC6-64, RC6-128, RC6-24, RC6-80)
• Round-trip tests for all standard word sizes
• Additional tests for 192/256-bit keys and CTR mode

Implementation Notes

• Pure JavaScript implementation using BigInt for arbitrary word sizes
• Supports non-power-of-2 word sizes per IETF specification
• Consistent API with existing CyberChef cipher operations (SM4, Blowfish, GOST, etc.)

References

• RC6 Block Cipher (Wikipedia)
• IETF Test Vectors for RC6 and RC5
• The RC6 Block Cipher (Original Paper)

[GCHQDeveloper581]

Is it worth (?possible?) making the number of rounds a configurable parameter rather than hard-coding it to 20?

(this confused me when I tried some of the other IETF test vectors and the output didn't match until I realised they were for a different number of rounds)

[thomasxm]

To make the number of rounds configurable rather than using the default and to match all the test vectors, we’d also need configurable word size. That would take me another few days, but it’s probably worth it to support the full parameterisation.

[thomasxm]

Done

[thomasxm]

Is it worth (?possible?) making the number of rounds a configurable parameter rather than hard-coding it to 20?

(this confused me when I tried some of the other IETF test vectors and the output didn't match until I realised they were for a different number of rounds)

When you have a chance, please review the new implementation :)