Match the vanilla client in the login/configuration handshake: cookies, brand, per-login chat session, known packs

[u9g]

Audit of nmp's client against the vanilla 1.21.4 client (ClientHandshakePacketListenerImpl, ClientCommonPacketListenerImpl, ClientPacketListener, KnownPacksManager, Options) turned up four places where nmp diverges in ways that servers and proxies notice. One commit each:

• Answer cookie_request (1.20.5+). nmp never replied, so any server or proxy that sends one (Velocity/Bungee modern forwarding, transfer flows) hangs the connection. A new src/client/cookies.js keeps client._cookies, fills it from store_cookie, and answers cookie_request in login, configuration and play with the stored value or an absent one, exactly like vanilla. A cookies option seeds the map so a consumer following a transfer packet can carry cookies over like vanilla's TransferState.
• Brand then Client Information, once. Vanilla answers login_acknowledged with custom_payload minecraft:brand followed by Client Information, and does not repeat either when a server sends it back to configuration (only configuration_acknowledged). nmp sent no brand, re-sent the settings on every configuration entry, and used view distance 10 where vanilla's default is 12. New brand option (default 'vanilla'; mineflayer's options.brand flows through), clientSettings defaults are now the vanilla ones.
• Fresh chat session on every login packet. After a proxy server switch the second login packet left the old session in place; vanilla discards it and announces a new session (new UUID, index 0) with the same key pair. client.once('login') → client.on('login').
• select_known_packs. Vanilla replies with the packs from the server's list it has locally. New knownPacks option ([{ namespace, id, version }]) answered with the intersection; the default stays [] so servers keep sending full registry data, which mineflayer/prismarine-registry rely on.

Docs updated in docs/API.md. In-process tests added in test/serverTest.js for cookie replies in all three states (including a stored-cookie echo and an absent reply), brand-then-settings sent exactly once across a start_configuration round trip, and the known-packs intersection.

Note: minecraft-data's 1.21.8 protocol.json declares cookie_response.value as a non-optional ByteArray (the 1.21.8 client writes it nullable like every other version). The absent reply is still wire-correct there (an empty ByteArray and an absent option both serialize as a single 0x00), but echoing a stored cookie needs a minecraft-data fix; the stored-cookie test is gated on the schema having the option wrapper.