Refresh settings environment schema

[gabsprogrammer]

Summary

Updates the Settings environment editor coverage so newly added DreamServer runtime variables are visible, documented, validated, and mapped to the correct apply/recreate targets.

Changes

• Adds missing .env.schema.json entries for Hermes/WhatsApp, APE, Dream Proxy, Host Agent, OpenClaw, Dream Talk voice URLs, Open WebUI STT/TTS settings, and llama.cpp tuning knobs documented in .env.example.
• Keeps .env.example aligned with the schema so documented keys appear in the dashboard Settings editor instead of falling through as unknown/advanced overrides.
• Maps WhatsApp/Hermes, Dream Proxy, OpenClaw, and APE env changes to the right service recreation plan.
• Makes Open WebUI STT/TTS compose values respect AUDIO_STT_* and AUDIO_TTS_* overrides from .env.
• Lets Dream Talk fall back from AUDIO_TTS_VOICE to legacy TTS_VOICE.
• Adds regression tests ensuring documented .env.example keys exist in the schema and apply-plan routing covers the new agent/proxy settings.

Validation

• python -m pytest dream-server/extensions/services/dashboard-api/tests/test_settings_env.py
• docker compose --env-file .env.example -f docker-compose.base.yml config --quiet
• git diff --check
• JSON parse check for .env.schema.json