How does ECB (Electronic Code Book) mode work, and what are its three serious weaknesses?

Each plaintext block is encrypted independently with the same key — there is no chaining between blocks.

Encryption: c_i = E(m_i, k) for each block, in parallel.

The weaknesses:

1. Identical plaintext blocks produce identical ciphertext blocks. This famously turns the "ECB penguin" — encrypting the Linux mascot image with AES-ECB leaves the penguin clearly visible because pixel blocks repeat. Any plaintext with structure (English text, JSON, images) leaks heavily.
2. Blocks can be swapped, removed, or duplicated unnoticed. An attacker who sees c_1 c_2 c_3 c_4 can rearrange to c_3 c_4 c_1 c_2 and the receiver decrypts to a valid (but reordered) message. Or simply delete c_2 — the rest still decrypts.
3. Partial encryption / decryption. Because blocks are independent, an attacker can target individual blocks. There's no integrity binding the whole message together.

Why it's still in textbooks: as the simplest mode to understand, and as the negative example. Adobe leaked 130M passwords in 2013 by storing them with 3DES in ECB mode — researchers built crossword-style attacks because identical passwords gave identical ciphertexts, instantly revealing user clusters and (via password hints) actual passwords.

Tip: Search "ECB penguin" for the most viral cryptography image of all time. If you remember nothing else, remember "don't use ECB."