What are the best dashboard skills for Claude Code?

Dashboards are where agent-generated UI usually falls apart: too many charts, no metric hierarchy, default-template styling. The catalog has a deep bench for exactly this, led by firecrawl-dashboard-reporting at 28,490 installs (skills.sh registry).

Key takeaways

• Dashboard skills split into three layers: designing the metrics, building the UI, and wiring observability; installing one of each beats installing three from the same layer.
• firecrawl-dashboard-reporting is the most-installed dashboard skill at 28,490 installs (skills.sh registry), automating web-data reporting dashboards.
• For dashboard design proper, kpi-dashboard-design (11,237 installs) from the wshobson/agents collection (37,769 GitHub stars) is the strongest pick.
• Observability is vendor-led: grafana-dashboards records 9,770 installs, and Grafana and Elastic both ship first-party skills.
• The wider field is ranked at best skills for frontend design and in the UI design hub.

Why do agent-built dashboards need a dedicated skill?

Because a dashboard is a design problem before it is a coding problem, and agents skip the design step. Without guidance, Claude Code renders every available metric at equal weight, which produces noise. A dashboard skill encodes the missing decisions: which KPIs earn the top row, what comparison each chart makes, when a table beats a graph. The catalog's build/frontend cluster holds 6,010 skills (Skillselion catalog, July 2026), but the dashboard-specific ones concentrate that judgment. The payoff compounds: a dashboard is usually the most-viewed screen a team ships internally, so a wrong hierarchy gets institutionalized fast.

What is the best skill for designing a KPI dashboard?

kpi-dashboard-design is the top pick at 11,237 installs, from the wshobson/agents collection with 37,769 GitHub stars (skills.sh registry, GitHub). It works upstream of code: metric selection, visual hierarchy, drill-down structure, and refresh cadence, so the agent produces a defensible layout instead of a chart dump.

npx skills add wshobson/agents

Strong alternatives: build-dashboard from anthropics (5,854 installs, 22,524 GitHub stars) turns data plus intent into a finished interactive artifact, and dashboard-builder (3,746 installs) is the general-purpose build skill from the everything-claude-code kit, whose repo carries 228,290 GitHub stars. The rest of the field is ranked at best skills for frontend design.

What is the most-installed dashboard skill overall?

firecrawl-dashboard-reporting at 28,490 installs (skills.sh registry). It occupies a different niche than the design skills: it crawls live web data and assembles recurring reporting dashboards, which makes it the pick when the dashboard is an output you deliver on schedule rather than a product surface you design once.

Which skills cover Grafana, Kibana, and monitoring dashboards?

Observability dashboards have their own conventions, and the vendor-adjacent skills encode them. grafana-dashboards (9,770 installs) covers panel design and PromQL-backed layouts; Grafana's first-party dashboarding skill (2,297 installs) reflects how the vendor itself structures boards; kibana-dashboards from elastic (1,956 installs) does the same for the ELK side, and building-dashboards from axiomhq adds 1,720 installs (all skills.sh registry). First-party skills are worth preferring here: the vendor knows its own query language's sharp edges.

What about admin panels and product dashboards?

Two skills carry this use case. building-admin-dashboard-customizations (2,649 installs, skills.sh registry) is medusajs's first-party skill for extending commerce admin UIs, and metrics-dashboard (1,741 installs, from a repo with 23,463 GitHub stars) approaches dashboards from the product-management side: which numbers a team should stare at weekly. Pair either with a general UI-quality skill from the frontend and design category so the output escapes the default admin-template look.

A complete starter stack

• kpi-dashboard-design (11,237 installs): decide what the dashboard says before building it.
• build-dashboard (5,854 installs): anthropics' skill for producing the working artifact.
• grafana-dashboards (9,770 installs): add when observability boards enter the picture.
• firecrawl-dashboard-reporting (28,490 installs): add for scheduled, web-data-driven reporting.

Common pitfalls

• Skipping the design layer: generating charts before deciding metric hierarchy guarantees a rebuild; run kpi-dashboard-design first.
• Using a generic UI skill for Grafana or Kibana work: panel JSON and query languages are vendor-specific, and the first-party skills exist because of it.
• Cramming every metric in: the skill can propose hierarchy, but keep the top row to three to five KPIs and push the rest behind drill-downs.

Design the metrics, build the surface, wire the monitoring: one skill per layer and the dashboard holds up.

Common questions