Summary
Build a web dashboard that visualizes ensemble run results:
- Side-by-side diff comparison across agents
- Convergence heatmaps showing which agents clustered together
- Historical trends (pass rate, convergence over time)
Why
The CLI output (compare, stats, evaluate) works but visual comparison of diffs and convergence patterns would make the ensemble results more accessible, especially for teams.