perf(vscode): speed up large session rendering

[marius-kilocode]

Summary

• Load session transcripts incrementally and virtualize turn rendering so large conversations open at the latest turn without mounting full history.
• Defer expensive historical markdown/tool-output work and lazy-load slash commands to reduce webview parse/render time during session switches.
• Reduce payload and first-paint markdown work for large historical tool outputs.

Closes #8607

[kilo-code-bot]

Code Review Summary

Status: No Issues Found | Recommendation: Merge

Files Reviewed (1 files)

• bun.lock

---

_{Reviewed by gpt-5.4-20260305 · 1,489,227 tokens}