[WASM] Integration tests: Conversation managers

[awsarron]

Parent issue

strands-agents/sdk-typescript#1104

Scope

Write integration tests validating conversation manager configurations:

• NullConversationManager() - no context modification
• SlidingWindowConversationManager(window_size=N) - fixed window trimming
• SummarizingConversationManager(...) - LLM-based summarization on overflow

Context

Conversation managers are config-only dataclasses on the Python side. The actual logic runs inside the TS SDK within the WASM guest. Tests need to verify:

• Config serializes correctly through the WIT boundary
• Sliding window trims messages when context grows large
• Summarizing manager produces summaries (requires multi-turn conversation that overflows)
• NullConversationManager leaves messages untouched

Done when

• NullConversationManager passes messages through unchanged
• SlidingWindowConversationManager trims old messages correctly
• SummarizingConversationManager summarizes on overflow
• Multi-turn conversations work correctly with each manager