Agent Dashboard
The Agent Dashboard is an interactive TUI that shows you exactly what your AI agents are doing — every command, its risk level, exit status, and timing — in a single real-time view.
Opening the Dashboard
suv agent dashboardThis opens the interactive terminal UI. The dashboard requires that agent tracking is set up for at least one agent. If no agent commands have been recorded yet, the dashboard will be empty.
Flags
| Flag | Description |
|---|---|
--after <TIME> | Show commands after a specific time (e.g., "3 days ago", "2025-01-15") |
--executor <NAME> | Filter to a specific agent (e.g., claude-code, cursor) |
--here | Only show commands from the current directory |
What the Dashboard Shows
Summary Panel
At the top of the dashboard, a summary panel displays aggregate metrics for the selected time period:
- Agent count — how many distinct agents have run commands
- Total commands — the total number of agent-executed commands
- Success rate — percentage of commands that exited with code 0
- Risk counts — breakdown of commands by risk level (Critical, High, Medium, Low, Safe)
Command Timeline
Below the summary, a chronological timeline lists every agent command. Each entry shows:
- Timestamp — when the command was executed
- Risk level indicator — a visual marker showing the assessed risk (
!!for Critical/High,~for Medium,.for Low,okfor Safe) - Exit status — success (green) or failure (red) with the exit code
- Command text — the full command that was executed
- Executor — which agent ran it
- Directory — the working directory at time of execution
Keyboard Shortcuts
The dashboard is fully keyboard-driven. Use these shortcuts to navigate and filter:
| Key | Action |
|---|---|
Up / Down | Navigate through the command timeline |
Tab | Toggle the detail pane for the selected command |
1 | Set time period to Today |
2 | Set time period to Last 7 days |
3 | Set time period to Last 30 days |
4 | Set time period to All time |
a | Cycle through agent filter (show all, or filter to a specific agent) |
r | Toggle risk-only filter (show only Medium and above risk commands) |
p | Open the Prompt Explorer |
Ctrl+Y | Copy the selected command to clipboard |
q / Esc | Quit the dashboard |
Examples
Monitor recent Claude Code activity
suv agent dashboard --after "3 days ago" --executor claude-codeOpens the dashboard filtered to Claude Code commands from the last 3 days. Use the r key to quickly filter down to only risky commands.
Check agent activity in the current project
suv agent dashboard --hereShows only agent commands that were executed in the current working directory and its subdirectories.
Review all agent activity today
suv agent dashboardOpens the dashboard with the default time period. Press 1 to switch to Today, or 2/3/4 to expand the time range.
p in the dashboard to jump directly into the Prompt Explorer and see which prompts triggered the commands you're looking at. This is especially useful for understanding why an agent ran a particular sequence of commands.