core: model provider config and runtime messages

[PerishCode]

Summary

• add TOML-backed provider config and share Chat Completions providers across DeepSeek/SiliconFlow
• model Santi workspace URIs, soul identity, <santi-system> session messages, and aborted partial assistant output
• update transcript/runtime projection so long provider first-token waits still show in-session feedback

Validation

• cargo fmt --all --check
• cargo test --workspace
• cargo clippy --workspace --all-targets -- -D warnings
• pnpm -r --if-present lint
• pnpm -r --if-present build
• flavor check --root . --config flavor.toml