LLM: add Volcengine and BytePlus presets

[Maaannnn]

Summary

• add 4-provider support for volcengine, volcengine-coding-plan, byteplus, and byteplus-coding-plan
• wire the new presets into researchclaw init, the quickstart wizard, and the example config
• document the new provider/env-var options in the English and Chinese READMEs

Behavior Changes

• researchclaw init now shows the four new providers after minimax, using the canonical base URLs and model catalogs from the provider contract
• the quickstart wizard now offers the same provider family and fills in base URL, API key env var, primary model, and fallback models for named providers
• LLMClient.from_rc_config() now resolves the Volcengine and BytePlus presets without requiring manual base URL entry

Codebase Search

• searched for provider registration and config entrypoints with rg before editing
• confirmed this repo uses preset-driven LLM config instead of a registry/plugin provider system

Tests

• .venv/bin/pytest tests/test_rc_cli.py tests/test_volcano_provider.py tests/test_minimax_provider.py
• .venv/bin/python -m py_compile researchclaw/llm/__init__.py researchclaw/cli.py researchclaw/wizard/quickstart.py tests/test_rc_cli.py tests/test_volcano_provider.py

Risks

• only the main README and docs/README_CN.md were updated; the other translated READMEs still mention the older provider list
• no live provider API calls were run because this branch does not include real provider credentials