Compile native LLM requests in session tests

[kitlangton]

Summary

• register native LLM provider routes from the session request adapter
• add a focused LLMClient.prepare test that compiles session-native requests through OpenAI Responses lowering
• keep this offline: no provider calls or runtime switching

Test Plan

• bun run test -- test/session/llm-native.test.ts test/session/llm.test.ts --test-name-pattern native
• bun typecheck
• bunx oxlint packages/opencode/src/session/llm.ts packages/opencode/src/session/llm-native.ts packages/opencode/test/session/llm.test.ts packages/opencode/test/session/llm-native.test.ts
• git diff --check

Stack

1. Consume native LLM events in session processing #26639
2. Add native LLM request adapter #26941
3. Compile native LLM requests in session tests #26946 👈 current
4. Add native OpenAI runtime opt-in #26947