feat(minimax): add voice cloning

[octo-patch]

Reason: Extend the existing MiniMax audio-generation integration with clone-audio upload and voice-clone response handling; current source and open TTS PR #2238 cover synthesis only.

Summary:

• Add feature-gated clone-audio upload with supported format validation and multipart request handling.
• Add voice cloning requests with required file, voice, and model fields plus structured success responses.
• Preserve provider error envelopes and cover both request boundaries with focused unit tests.

Checks:

• cargo fmt --manifest-path repos/0xPlaygrounds__rig__recvrq1ox3FZrs/Cargo.toml --all -- --check
• cargo test --manifest-path repos/0xPlaygrounds__rig__recvrq1ox3FZrs/Cargo.toml -p rig-core --features audio providers::minimax --lib
• cargo clippy --manifest-path repos/0xPlaygrounds__rig__recvrq1ox3FZrs/Cargo.toml -p rig-core --features audio --lib --tests -- -D warnings -A clippy::useless-conversion -A clippy::collapsible-match -A clippy::unnecessary-sort-by
• cargo doc --manifest-path repos/0xPlaygrounds__rig__recvrq1ox3FZrs/Cargo.toml -p rig-core --features audio --no-deps