Add end-to-end encrypted voice messages; release v5.4.5

- Record voice notes in-browser, sent over the chunked AES-GCM file-transfer
  channel (per-file session key + signed SHA-256 integrity).
- Captured as PCM and encoded to WAV for universal playback (incl. iOS/Safari);
  auto-accepted and played inline from an in-memory blob, never written to disk.
- Composer mic button with live waveform + timer; desktop shows mic + send side
  by side, mobile swaps mic to send when typing.
- CSP media-src now allows blob: so recorded/received audio can play.
- Roadmap: Desktop Edition -> 5.0, new 5.5 'Secure Voice & Calls', later
  milestones shifted; version bumped to 5.4.5.
- Update README, docs (security/API/cryptography), and CHANGELOG.

doc/CRYPTOGRAPHY.md

@@ -2,7 +2,7 @@
 
 ## Release context
 
-- Product release: `v4.8.5`
+- Product release: `v5.4.5`
 - Protocol version: `4.1`
 
 ## Session establishment
@@ -16,6 +16,9 @@ The SAS is deterministic for both peers in the same authenticated session. Users
 - encrypted payloads are validated before decryption
 - decrypted chat text is sanitized before entering React state or the UI
 - replay and ordering controls remain part of the session layer
+- voice messages are transported over the file-transfer channel: each is
+  encrypted with a per-file AES-GCM session key and integrity-checked with a
+  signed SHA-256 hash before playback
 
 ## Local key metadata
 
@@ -23,4 +26,4 @@ Sensitive IndexedDB metadata is stored in encrypted envelopes. Legacy plaintext
 
 ## Scope note
 
-This document describes the current browser implementation behavior relevant to the v4.8.5 hardening release. It does not replace independent cryptographic review.
+This document describes the current browser implementation behavior relevant to the v5.4.5 release. It does not replace independent cryptographic review.