main
Add eval sets, sessions, and per-session matrix view
Remove Pairwise and Single Model eval modes
Remove unused formatTimeAgo function
Support text-brief eval sets alongside image sets
Add side-by-side compare viewer for eval outputs
Add rerun-on-failure support
add caching-related tools and remove prompt_cache_key
Add full agent-run logging with SQLite index and /evals/agent-runs viewer
rate on multiple axis
Batch image editing and background removal tools in parallel