[terence71-glitch]

Summary

• Avoids an unnecessary deep clone of Claude-native messages during Claude Code semantic passthrough preparation.
• Keeps the downstream final message-cloning path intact, preserving request output isolation while reducing memory/CPU overhead for large Claude Code histories.

Context

This fix is based on a code review suggestion from @gemini-code-assist[bot] in PR #2351

prepareClaudeCodeCompatibleSemanticBody() was deep-cloning claudeBody.messages, and buildClaudeCodeCompatibleRequest() then passed those messages through cloneClaudeCodeCompatibleMessagesFromClaude(), which clones the message objects again.
This removes the first clone because the final cloning step still protects the emitted request body.

Validation

• node --import tsx/esm --test tests/unit/claude-code-compatible-request.test.ts — passed (10/10)
• git diff --check — passed
• node --import tsx/esm --test tests/integration/*.test.ts — still fails with the same pre-existing integration failures observed before the fix

[gemini-code-assist]

Code Review

This pull request modifies the prepareClaudeCodeCompatibleSemanticBody function in open-sse/services/claudeCodeCompatible.ts to stop cloning the messages array from the input object, opting instead to cast it directly. This change likely improves performance by avoiding an unnecessary deep clone operation during the preparation of the semantic body. I have no feedback to provide as there were no review comments to evaluate.

[diegosouzapw]

Thanks @terence71-glitch! Excellent optimization. Merged into the upcoming v3.8.0 release.