feat: add Piper TTS provider

[followsound]

Summary

• add a local Piper TTS provider for ONNX voice models
• expose Piper in TTS setup/status and docs
• add focused Piper provider tests

Test plan

• python -m py_compile tools/tts_tool.py hermes_cli/setup.py
• python -m pytest tests/tools/test_tts_piper.py tests/tools/test_tts_kittentts.py -q -o 'addopts='\n\n## Notes\nThis enables voices such as the BMO/Bemo Piper model from https://github.com/brenpoly/be-more-agent/releases/tag/v1.0-voice via config:\n\nyaml\ntts:\n provider: piper\n piper:\n model: ~/.hermes/voices/bmo/bmo.onnx\n config: ~/.hermes/voices/bmo/bmo.onnx.json\n

[alt-glitch]

This was generated by AI during triage.

Native Piper TTS already shipped in main via the merged #17885 (closes #8508): tools/tts_tool.py has _import_piper() / _generate_piper_tts / _resolve_piper_voice_path, and the built-in resolver already accepts an absolute path to a user-supplied .onnx file — which covers the BMO/Bemo voice from your config example. Set:

tts:
  provider: piper
  piper:
    voice: ~/.hermes/voices/bmo/bmo.onnx

This PR's branch appears to predate that merge (its diff still describes "seven" providers, but current main has ten including Piper), so it re-adds an executable-based Piper that duplicates the shipped native provider. Marking as a duplicate of #17885.

[followsound]

Apologies was out of sync!