dashboard

auto-generated	true
name	dashboard
description	Project orientation, status overview, and skill/agent discovery. Shows metrics, blockers, available skills, agents, and quick-start actions.

Overview

The Dashboard skill gives you a comprehensive view of the project's current state: available skills and agents, task metrics, blockers, and suggested next actions.

Invoke with /dashboard.

Step 1: Available Skills

Present all available skills, grouped by scope:

Project Skills (from @tasks/process/)

Skill	Command	Purpose
Multi-File Changes	/multi-file-changes	Methodology for 3+ file changes, vendored code, cross-crate work
Task Runner	/task	Pick and start the next task from @tasks
Dashboard	/dashboard	This skill - orientation and project overview

Global Skills (from ~/.claude/)

Skill	Command	Purpose
Architect	/architect	Create architectural plans with alternatives analysis and edge cases
Feature Dev	/feature-dev	Guided feature development with codebase understanding
Find Skills	/find-skills	Discover and install new agent skills
Veo Video Generation	/veo-video-generation	Generate videos from images or text using Google Veo API

To discover more installable skills, use /find-skills.

Step 2: Available Agents

Present all 5 project agents with their domains and assignment guidance:

Agent	Domain	Assign When
svelte-developer	Svelte 5 UI, runes, Tauri IPC	Pages, components, stores, frontend IPC
rust-architect	Tauri commands, module design, async	Backend commands, module architecture, traits
testing-engineer	Vitest, Playwright, integration	Unit tests, E2E tests, test infrastructure
lexical-specialist	Lexical editor, custom nodes, Markdown	Editor plugins, custom nodes, transformers
writing-app-designer	Feature specs, UX, entity schemas	Specifications, UX flows, data models, file formats

Agent definitions live in @tasks/agents/*.org. For the full assignment matrix, see Agent Registry: Assignment Matrix.

Step 3: Task Metrics

Query the PRD system for current metrics:

emacsclient -s sakya -e '(prd-dashboard-cli)'

Present the metrics clearly:

Overall Progress

Show totals from the dashboard JSON:

• Total tasks, complete, in-progress, blocked, pending
• Overall completion percentage

Progress by Category

For each category (PROJ-001 through PROJ-006, any BUG-, IMP-):

• Title, total items, completed items, progress bar/percentage

Agent Workload

From the agents section of the dashboard JSON:

• Tasks assigned vs completed per agent
• Which agents have the most pending work

Error Handling

If emacsclient is unavailable, fall back to direct file scanning:

1. Read all @tasks/{projects,bugfixes,improvements}/*.org files
2. Count TODO keywords: ITEM, DOING, REVIEW, DONE, BLOCKED
3. Group counts by parent file
4. Present manually-computed metrics

Step 4: Blocked Items

Query for blocked tasks:

emacsclient -s sakya -e '(prd-list-blocked)'

For each blocked item, show:

• The blocked item's ID and title
• What it's blocked by (dependency IDs and their current status)
• Whether the blocker is close to completion (DOING/REVIEW)

Highlight critical blockers: items that block many downstream tasks.

Step 5: Recent Completions

Query velocity data:

emacsclient -s sakya -e '(prd-velocity-report 7)'

Show:

• Tasks completed in the last 7 days
• Velocity trend (increasing, stable, decreasing)
• Which categories had the most activity

Step 6: Quick-Start Actions

Based on the current state, suggest contextual next actions:

If there are blocked items with resolvable blockers

Suggest: "Unblock ITEM-XXX by completing ITEM-YYY (currently DOING, agent: rust-architect)"

If there are high-priority pending items

Suggest: "Start on ITEM-XXX (Priority #A, effort: 1h) — use /task ITEM-XXX"

If a project is near completion

Suggest: "PROJ-XXX is 90% complete — 2 items remaining to finish"

If no tasks are available or all are blocked

Suggest: "All tasks are blocked or complete. Consider:"

• Creating new tasks for upcoming work
• Reviewing BLOCKED items for stale dependencies
• Running /find-skills to discover new capabilities

General suggestions

• "Run /task to pick your next task"
• "Run emacsclient -s sakya -e '(prd-validate-all-cli)' to check for validation issues"

See Also

• Process Registry
• Task Runner - For picking and starting a specific task
• Multi-File Changes - For complex cross-file work
• Agent Registry - All available agents
• CLAUDE.md - Claude Code integration guide
• Reference Guide - Full PRD system documentation