Overview
{"content":"Valdr's UI is your operating surface for agent-driven delivery. Four dashboard tabs give you instant signal on project health, agent workload, and token cost. Tasks carry specs and acceptance criteria so nothing ships without verification. Plans generate structured work, sprints time-box it, and the orchestrator coordinates agents across all of it.\n\n{{\u003c thumbcard src=\"/images/ui/dashboard/workspace-pulse.png\" alt=\"Workspace Pulse dashboard showing KPI cards, throughput trend, and perspective carousel\" caption=\"Workspace Pulse — KPI cards, throughput trends, and actionable perspectives at a glance\" \u003e}}\n\nOpen the Dashboard and you're looking at live KPI cards, trend charts, and a perspective carousel that surfaces the next thing to act on — all scoped to the time range you choose.\n\n{{\u003c callout type=\"info\" \u003e}}\n**The Dashboard is where you decide, not just observe.** Every KPI card is clickable, every perspective links to the relevant detail view, and every metric updates with the selected time range.\n{{\u003c /callout \u003e}}\n\n## What this page covers\n\n- Sidebar navigation and the core screens\n- How the four dashboard tabs map to different operational questions\n- A fast triage workflow for daily check-ins\n\n## Navigation at a glance\n\n| Screen | What it does |\n|--------|-------------|\n| **Dashboard** | Four tabs: [Workspace Pulse](/valdr/docs/ui/dashboard/workspace-pulse/) for delivery flow, [Project Breakdown](/valdr/docs/ui/dashboard/project-breakdown/) for per-project comparison, [Agent Workload Pulse](/valdr/docs/ui/dashboard/agent-workload-pulse/) for quality and productivity, and [Agent Drilldown](/valdr/docs/ui/dashboard/agent-drilldown/) for run-level token, cost, and reliability analysis. |\n| **Projects** | Status, owners, and linked tasks for every project. Spot stalled projects and ownership gaps. |\n| **Planner** | Shape new work while seeing existing load. Generate plans from natural language descriptions. |\n| **Plans** | Structured feature plans with requirements, specs, and task generation. Plans are the source of truth for what to build. |\n| **Sprints** | Time-boxed delivery cycles with task backlogs, completion tracking, and status lifecycle. |\n| **Tasks** | Full task lifecycle from backlog to done, with acceptance checklists, agent assignment, and review gating. |\n| **Agents** | Agent registry with capabilities, prompts, and session history. Configure what each agent can do. |\n| **Prompts** | Searchable catalog of 167+ instruction fragments with role classification, token sizing, and usage tracking. |\n| **Capabilities** | Reusable skill bundles that wire prompts to agents. |\n| **Agent Sessions** | Inspectable execution history for every agent run — tokens, duration, commands, and outcomes. |\n| **Orchestrator** | Multi-agent coordination and dispatch through the A2A mailbox. |\n| **Settings** | Workspace configuration and provider setup. |\n\n## Flow that reduces churn\n\n1. **Start on Dashboard** — check the Workspace Pulse for headline KPIs: open tasks, at-risk items, throughput, and sprint health.\n2. **Drill into projects** — switch to Project Breakdown to compare backlogs and completion rates across projects.\n3. **Check agent health** — use Agent Workload Pulse for quality scores and review activity, then Agent Drilldown for token cost and reliability.\n4. **Act on tasks** — open Tasks to reassign, unblock, or review the items flagged by dashboard perspectives.\n5. **Reinforce plans** — verify that Plans and acceptance criteria still match what's being executed.\n\n**When a signal shifts, act fast:**\n- At-risk items rising → click through to see stale reviews, blocked tasks, or approaching deadlines.\n- Token pressure high → open Agent Drilldown, identify the top consumer, and optimize prompts or switch models.\n- Score coverage low → increase audit frequency or enable automated scoring.\n- Unassigned tasks → assign owners before work stalls.\n\n## Fast triage workflow\n\n{{% steps %}}\n\n### Read the pulse\n\nCheck Dashboard → Workspace Pulse for KPI cards and the perspective carousel. Scan for non-zero at-risk counts, declining throughput, or growing backlogs.\n\n### Check agent cost and reliability\n\nSwitch to Agent Drilldown. Review failure rate, token pressure, and score coverage. If one agent is consuming 80% of tokens, that's where to optimize.\n\n### Focus on blockers\n\nBrowse the perspective carousel — Sprint Health, Review Queue, Stale In Progress, and Top Blockers surface the highest-leverage actions.\n\n### Drive execution\n\nOpen Tasks, filter by status or assignee, and clear the review queue first to unblock shipping.\n\n{{% /steps %}}\n\n## Why this matters\n\n- **Issues surface early**: KPI cards and perspectives flag risk before it compounds — stale reviews, unassigned work, and token budget spikes become visible in seconds.\n- **Predictable delivery**: plans, sprints, and tasks stay linked with acceptance criteria and review gating. Fewer surprises at ship time.\n- **Agent accountability**: every agent run is attributed, scored, and costed. You know exactly what your AI workforce is doing and whether it's worth it.\n- **Complete local audit trail**: every change, every run, every review is recorded in your workspace while the data exists.\n\nNext, dive into the [Dashboard](/docs/ui/dashboard/) to explore all four tabs in detail.","description":"Navigate Valdr's UI to manage projects, agents, and delivery from a single workspace.","title":"Overview"}
Valdr’s UI is your operating surface for agent-driven delivery. Four dashboard tabs give you instant signal on project health, agent workload, and token cost. Tasks carry specs and acceptance criteria so nothing ships without verification. Plans generate structured work, sprints time-box it, and the orchestrator coordinates agents across all of it.
Workspace Pulse — KPI cards, throughput trends, and actionable perspectives at a glance
Open the Dashboard and you’re looking at live KPI cards, trend charts, and a perspective carousel that surfaces the next thing to act on — all scoped to the time range you choose.
The Dashboard is where you decide, not just observe. Every KPI card is clickable, every perspective links to the relevant detail view, and every metric updates with the selected time range.
What this page covers
- Sidebar navigation and the core screens
- How the four dashboard tabs map to different operational questions
- A fast triage workflow for daily check-ins
Navigation at a glance
| Screen | What it does |
|---|---|
| Dashboard | Four tabs: Workspace Pulse for delivery flow, Project Breakdown for per-project comparison, Agent Workload Pulse for quality and productivity, and Agent Drilldown for run-level token, cost, and reliability analysis. |
| Projects | Status, owners, and linked tasks for every project. Spot stalled projects and ownership gaps. |
| Planner | Shape new work while seeing existing load. Generate plans from natural language descriptions. |
| Plans | Structured feature plans with requirements, specs, and task generation. Plans are the source of truth for what to build. |
| Sprints | Time-boxed delivery cycles with task backlogs, completion tracking, and status lifecycle. |
| Tasks | Full task lifecycle from backlog to done, with acceptance checklists, agent assignment, and review gating. |
| Agents | Agent registry with capabilities, prompts, and session history. Configure what each agent can do. |
| Prompts | Searchable catalog of 167+ instruction fragments with role classification, token sizing, and usage tracking. |
| Capabilities | Reusable skill bundles that wire prompts to agents. |
| Agent Sessions | Inspectable execution history for every agent run — tokens, duration, commands, and outcomes. |
| Orchestrator | Multi-agent coordination and dispatch through the A2A mailbox. |
| Settings | Workspace configuration and provider setup. |
Flow that reduces churn
- Start on Dashboard — check the Workspace Pulse for headline KPIs: open tasks, at-risk items, throughput, and sprint health.
- Drill into projects — switch to Project Breakdown to compare backlogs and completion rates across projects.
- Check agent health — use Agent Workload Pulse for quality scores and review activity, then Agent Drilldown for token cost and reliability.
- Act on tasks — open Tasks to reassign, unblock, or review the items flagged by dashboard perspectives.
- Reinforce plans — verify that Plans and acceptance criteria still match what’s being executed.
When a signal shifts, act fast:
- At-risk items rising → click through to see stale reviews, blocked tasks, or approaching deadlines.
- Token pressure high → open Agent Drilldown, identify the top consumer, and optimize prompts or switch models.
- Score coverage low → increase audit frequency or enable automated scoring.
- Unassigned tasks → assign owners before work stalls.
Fast triage workflow
Read the pulse
Check Dashboard → Workspace Pulse for KPI cards and the perspective carousel. Scan for non-zero at-risk counts, declining throughput, or growing backlogs.
Check agent cost and reliability
Switch to Agent Drilldown. Review failure rate, token pressure, and score coverage. If one agent is consuming 80% of tokens, that’s where to optimize.
Focus on blockers
Browse the perspective carousel — Sprint Health, Review Queue, Stale In Progress, and Top Blockers surface the highest-leverage actions.
Drive execution
Open Tasks, filter by status or assignee, and clear the review queue first to unblock shipping.
Why this matters
- Issues surface early: KPI cards and perspectives flag risk before it compounds — stale reviews, unassigned work, and token budget spikes become visible in seconds.
- Predictable delivery: plans, sprints, and tasks stay linked with acceptance criteria and review gating. Fewer surprises at ship time.
- Agent accountability: every agent run is attributed, scored, and costed. You know exactly what your AI workforce is doing and whether it’s worth it.
- Complete local audit trail: every change, every run, every review is recorded in your workspace while the data exists.
Next, dive into the Dashboard to explore all four tabs in detail.