← ClaudeAtlas

dashboard-designlisted

Anti-slop dashboard skill for SaaS app UI. Admin panels, analytics, CRM, data tables, app shells that read as real daily-use software. Project-first component retrieval, strict visual QA, zero placeholder data.
mridulthareja/dashboard-skill · ★ 0 · Web & Frontend · score 67
Install: claude install-skill mridulthareja/dashboard-skill
# dashboard-design: Anti-Slop Dashboard Skill > SaaS dashboards, admin panels, analytics, CRM, data-dense app UI. Not landing pages, not marketing sections, not Dribbble concepts. > Every rule below is **contextual**. None of it fires automatically. First read the product, then pull only what fits. --- ## 0. BRIEF INFERENCE (Read the Product Before Anything Else) Before touching code or picking components, **infer what the product actually is**. Most LLM dashboard output is bad because the model jumps to a default layout (KPI row + chart + table) instead of reading the room. ### 0.A Read these signals first 1. **Product kind** - CRM / admin / analytics / monitoring / billing / settings / AI console / trading terminal / support desk / file manager / audit log. 2. **Primary user + job** - role doing what task, at what frequency. Daily operator vs. occasional admin picks the density. 3. **Core workflow** - the 3–5 step loop they repeat. The UI exists to serve this loop, nothing else. 4. **Entities** - 3–6 nouns + the 2–4 fields that matter for each. If you cannot name them, you cannot lay out the page. 5. **Primary actions** - verbs: create, approve, assign, export, retry, escalate. Exactly one primary action per view. 6. **Brand assets that already exist** - tokens, component library, icon set, chart stack. For existing codebases these are starting material, not optional input (see Section 11). 7. **Quiet constraints** - regulated industries, accessibility-first audiences