dsh-whale-tts: Give DSH Whale a Voice

Maintained by cilis, this voice synthesis plugin for DeepSeek Harness (DSH) is licensed under MIT. It synthesizes text into local audio files and returns the absolute path, relative path, and filename, leaving playback, sending, and uploading to the caller. It provides three tools: `tts_synthesize` for text-to-speech, `tts_list_voices` to list provider capabilities and saved voice profiles, and `tts_status` for self-checking the provider, key, cache, and output directory. It supports voice design and cloning: descriptions generate 2–5 candidates, or wav/mp3 files within 10MB can be uploaded for reference audio cloning, with tone and speed retained as style instructions. Voice profiles are categorized into presets, cloning, and design, with a maximum of 50 profiles. The DSH plugin page offers bilingual (Chinese/English) settings cards, allowing users to listen to voices, manage profiles, enter keys, and tune output. It also enables auto-reading of new replies (default is off, reading only the last body text of a single turn). It supports exporting JSON backups without keys, as well as merge or replace imports. Keys are stored only as reference names and do not appear in tool results, logs, or the frontend. The main provider is currently MiMo, which falls back to Windows offline SAPI when unavailable. Installation is performed via `dsh plugin --profile <profile> add file:/path/to/dsh-whale-tts-0.5.0.tgz`, or by installing a tgz/local directory from the plugin page followed by a restart; it requires DSH >= 0.2.1-rc.4 < 0.3.0-0 and Node >= 24.9.0. The plugin runs with the permissions of the current DSH process; users should check the source code and license before installation.