Classical Cipher Workbench

Classical ciphers are the substitution schemes that predate modern cryptography. They are useless for real secrecy but everywhere in CTFs, puzzles, and CT-style challenges, and they are a superb way to see why cryptanalysis works. This workbench applies the four most common transforms at once and profiles the letter frequencies of your text.

The ciphers. A Caesar shift rotates each letter by a fixed amount; ROT13 is just Caesar with a shift of 13, and because 13 is half of 26 it is its own inverse. Atbash mirrors the alphabet (A↔Z, B↔Y) and is likewise self-inverse. Vigenère uses a repeating keyword to apply many different Caesar shifts, which defeats naive single-shift frequency analysis — but only up to a point.

Frequency analysis. English text has a very uneven letter distribution (E, T, A, O dominate), and the tool reports the most common letters plus the Index of Coincidence (IoC). A monoalphabetic cipher preserves the plaintext's IoC (~0.067 for English), so a high IoC says "try Caesar/Atbash/substitution." A Vigenère cipher flattens the distribution toward random (~0.038), and a low IoC is the signal to hunt for the key length (via Kasiski examination or the Friedman test) and then break each column as a separate Caesar.

Worked example. To brute-force a Caesar you only have 25 shifts to try — step the shift value and watch for readable output. For Vigenère, once you know the key length, each set of every-Nth letter is a simple Caesar you can solve by frequency.

Defensive angle. The lesson these ciphers teach is exactly why we no longer use them: any scheme where structure of the plaintext survives into the ciphertext leaks information. Modern ciphers (AES, ChaCha20) are designed so output is statistically indistinguishable from random. If you ever find "encryption" in an application that is really ROT13, a fixed-byte XOR, or Base64, treat it as plaintext — it protects nothing. When a CTF string is not a classical cipher at all, try decoding it as Base64 or hex first, and check its shape against known file signatures — half of "crypto" challenges are really just layered encodings.