chore: skip tracking judge results that were not sampled

[jsonbailey]

Summary

• Skip processing judge results where r.sampled is False in ManagedModel._run_and_track
• Prevents unsampled evaluations from being tracked unnecessarily

Test plan

• Verify judge results with sampled=False are skipped without errors
• Verify judge results with sampled=True continue to be tracked as before

🤖 Generated with Claude Code

---

Note

Low Risk
Low risk: adds a simple sampled gate before tracking judge results, affecting only metrics/logging for evaluations that were intentionally skipped by sampling.

Overview
Judge result tracking now ignores unsampled evaluations. In ManagedModel and ManagedAgent, _run_and_track skips any JudgeResult where sampled is False, so only actually-executed judge runs are passed to tracker.track_judge_result (and related warnings).

^{Reviewed by Cursor Bugbot for commit e45e7f7. Bugbot is set up for automated code reviews on this repo. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, have a team admin enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 34624ba. Configure here.}