docs: clarify RunResult testing surface

[nightcityblade]

Fixes #5410

Summary

• document RunResult.events as the primary minimal testing-facing surface
• clarify that events stay typed and chronological for external consumers
• note that final_output is optional and best treated as secondary

Testing

• python3 -m compileall livekit-agents/livekit/agents/voice/run_result.py livekit-agents/livekit/agents/voice/agent_session.py

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.

---

nightcityblade seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[devin-ai-integration]

✅ Devin Review: No Issues Found

Devin Review analyzed this PR and found no potential bugs to report.

View in Devin Review to see 1 additional finding.

[theomonnom]

lgtm! but AgentSession.run is not only for testing. It can be used to build text apps

[nightcityblade]

Good call, thanks. I agree AgentSession.run() is broader than a testing-only entrypoint. I opened a narrower follow-up docs PR around the small testing surface on RunResult.events instead of framing run() itself as test-only.