AbstractVoice v0.10.2

Added

• Added model-aware STT discovery across the REPL and the AbstractCore plugin,
  including faster-whisper aliases such as large plus Transformers ASR ids
  such as openai/whisper-large-v3-turbo and Qwen/Qwen3-ASR-1.7B.
• Added provider-scoped STT catalog data in voice_catalog() via
  stt_models_by_provider and stt_engine_variants, and taught the
  AbstractCore plugin to accept combined selectors such as
  faster-whisper:large and transformers-asr:Qwen/Qwen3-ASR-1.7B.

Fixed

• Preserved the selected faster-whisper model across REPL VoiceManager
  recreation paths such as /tts on, TTS engine switching, and memory restore,
  and refreshed cached faster-whisper adapters when /whisper changes the
  model.

Changed

• Updated the README, API docs, REPL guide, llms.txt, llms-full.txt, and
  CLI/web help text to reflect the current STT provider/model abstraction and
  AbstractCore catalog surface.