Add GMCP support and shared media layer

README.md

@@ -11,6 +11,9 @@ The project is currently an early but usable Telnet client. Its interface is bui
 - UTF-8, Windows-1252, and ISO-8859-1 character encodings
 - Streaming Telnet negotiation with ECHO, Suppress Go Ahead, and Terminal Type support
 - MCCP2 compressed-stream negotiation and bounded streaming decompression
+- GMCP negotiation with bounded UTF-8/JSON parsing and a raw diagnostic inspector
+- Accessible live views for character, room, items, skills, group, and communication data
+- Opt-in Client.Media audio with captions, per-profile volumes, HTTPS-only bounded downloads, and cache limits
 - Streaming ANSI control-sequence filtering
 - Accessible scrollback and command entry
 - Automatic reading of incoming lines and prompts
@@ -40,6 +43,9 @@ The output and command fields provide additional actions through their context m
 - `lib/meow/profile_repository.rb` stores and validates connection profiles.
 - `lib/meow/transport.rb` owns TCP/TLS sockets and background I/O.
 - `lib/meow/telnet.rb` implements the streaming Telnet state machine.
+- `lib/meow/gmcp.rb` parses GMCP and maintains package state.
+- `lib/meow/media.rb` provides protocol-neutral media requests, safe caching, and Elten audio playback.
+- `lib/meow/session_gmcp.rb` integrates GMCP packages with accessible session views.
 - `lib/meow/text_pipeline.rb` handles ANSI filtering, decoding, lines, and prompts.
 - `lib/meow/session.rb` owns connected-session state, scrollback, input, and speech.
 - `lib/meow/ui.rb` implements the connection manager and profile editor.
@@ -52,11 +58,10 @@ Automated regression coverage now exercises the streaming Telnet parser and text
 
 Planned work, roughly in order:
 
-1. Add GMCP negotiation, bounded JSON parsing, typed package events, and an accessible inspector.
-2. Add opt-in MSP sound and music with safe URL handling, limits, caching, and independent playback controls.
-3. Add aliases, triggers, timers, optional transcripts, and configurable command shortcuts.
-4. Add multiple simultaneous sessions with independent output and speech queues.
-5. Add localization catalogs and package/signing automation for releases.
+1. Add MSP parsing as a second adapter to the shared media request and playback layer.
+2. Add aliases, triggers, timers, optional transcripts, and configurable command shortcuts.
+3. Add multiple simultaneous sessions with independent output, media, and speech queues.
+4. Add localization catalogs and package/signing automation for releases.
 
 ## Running the tests