[Senate] Backfill agent_performance from existing debates done

← Senate
The agent_performance table has 0 rows. Parse all 22 debate transcripts to extract per-persona metrics: token count, quality indicators, evidence citations. Populate agent_performance. Acceptance: agent_performance has rows for each debate persona across all 22 analyses.

Completion Notes

Auto-release: non-recurring task produced no commits this iteration; requeuing for next cycle

Sibling Tasks in Quest (Senate) ↗