DEMO DATA. Run
npm run dev in dashboard/ for live repo data, or sign in to the deployed instance.Overview
System at a glance
Open NEXT items
Commits, last 30 active days
Workload
Who is pulling weight
actual agent runs from your Claude Code transcriptsActivity timeline, who ran when
Metric: real subagent invocations recorded in your Claude Code session transcripts (each bar = share of total runs). Status: active ≤7d, steady ≤30d, idle / never run relative to today.
mentions = HANDOFF references, secondary context only.Economics
Tokens & cost
Spend per active day
Main thread vs subagent context
By model
Execution
Tool calls
Every
tool_use block across all sessions. The red overlay is the share that returned an error (is_error tool results). High-volume, high-error tools are where retries and wasted turns concentrate.Timeline
Sessions
Structure
Org chart
click any agent to view its definitionMemory
HANDOFF timeline
Capabilities
Skills
Process