fix: remove content rstrip in normalize_response to preserve TITO prefix match

[mikasenghaas]

Summary

• Removes the content.rstrip() call in normalize_response that was stripping trailing whitespace from message content, which broke TITO (think-in-turn-out) prefix matching downstream.

🤖 Generated with Claude Code

---

Note

Low Risk
Small, localized change to response normalization that only affects how assistant content is stored/compared; main risk is slightly different cache/prefix-hit behavior due to preserved trailing whitespace.

Overview
Stops stripping trailing whitespace from assistant message.content in OpenCodeEnv.normalize_response, so the stored trajectory matches the agent’s original output for TITO prefix cache hits.

Tool-call normalization (lowercasing tool names and compacting JSON arguments) remains unchanged; only the content trimming behavior is removed.

^{Written by Cursor Bugbot for commit 5197334. This will update automatically on new commits. Configure here.}