ARCHIVE DESK LOCKED FILE FORMAT V1

All offsets are bytes. Multi-byte integers are unsigned big-endian. The header is exactly48bytes.
0..7 ASCII ADLOCK01
8 algorithm ID1, AES-256-GCM
9 KDF ID1, PBKDF2-HMAC-SHA-256
10..13 iteration count600000, other values rejected
14..29 fresh random16byte salt
30..41 fresh random12byte GCM nonce
42..45 original ZIP byte length
46..47 reserved zero
48..end ciphertext followed by the16byte GCM authentication tag

The full48byte header is AES-GCM additional authenticated data. The ZIP bytes are the plaintext. There is no extra JSON metadata, password verifier or filename in the header.
The complete envelope must have48+ZIP_length+16bytes. ZIP_length must be nonzero and no more than52MiB. The parser rejects unsupported IDs, reserved fields, iteration counts, truncation, appended bytes and invalid lengths before deriving a key.
PBKDF2 receives the literal UTF-8 encoding of the passphrase. No trimming, case folding or Unicode normalization. Malformed Unicode, all-whitespace input, fewer than12Unicode code points or more than1024UTF-8bytes are rejected.
A successful GCM decrypt is followed by the existing strict Archive Desk ZIP parser, CRC and manifest SHA-256 checks. Authentication success alone does not admit arbitrary bytes as a valid backup.

The counter choice follows the600000iteration PBKDF2-HMAC-SHA256 guidance from OWASP's password-storage reference. It is a parameter choice, not a claim of FIPS validation or an encryption-system audit.
https://cheatsheetseries.owasp.org/cheatsheets/Password_Storage_Cheat_Sheet.html
GCM parameter reference:
https://developer.mozilla.org/en-US/docs/Web/API/AesGcmParams
