fix: preserve audio MCP content blocks

[Bortlesboat]

Summary

• preserve AudioContent in create_sdk_mcp_server instead of logging and skipping it
• preserve AudioContent in _handle_sdk_mcp_request so the JSON-RPC bridge passes it through to the CLI
• add regression coverage for both code paths

Context

PR #725 fixed resource_link and embedded resources, but audio was still dropped in both handlers because neither content-type switch handled it.

Test plan

• python -m pytest tests/test_sdk_mcp_integration.py
• python -m pytest tests/
• python -m ruff check src/ tests/
• python -m mypy src/

[Bortlesboat]

Quick follow-up from my side: I’m treating review-required PRs as top priority this week. If you want any specific changes, rebase, or split, I can turn them around quickly.

[Bortlesboat]

Rebased this onto current main and narrowed it to the remaining AudioContent gap that was left after #725.

Fresh verification on this branch:

• python -m pytest tests/test_sdk_mcp_integration.py
• python -m pytest tests/
• python -m ruff check src/ tests/
• python -m mypy src/