dashboard-intelligence

Coordinates governed dashboard design from approved report intent to human blueprint review.

2|Updated Jun 22, 2026
One-click install
npx skills add https://github.com/Kemetra/Seshat-BI --skill dashboard-intelligence-kemetra
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: dashboard-intelligence
Source: https://github.com/Kemetra/Seshat-BI/tree/main/.claude/skills/dashboard-intelligence
Command: npx skills add https://github.com/Kemetra/Seshat-BI --skill dashboard-intelligence-kemetra

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Dashboard work in the Seshat BI repo spans several individually-governed capabilities (gap checks, planning, design, composition, QA) that users must wire together by hand, making it possible to reach visual implementation without a committed report purpose or approved metric contracts. This Skill sequences those shipped capabilities into one fail-closed journey that stops at the human blueprint review. ## Core Features & Use Cases - State-driven coordination: Reads committed repository state via a read-only coordinator helper to pick exactly one next allowed action or a named blocker, with no separate state file. - Hard governance gates: Enforces semantic_model_ready: pass before any data-bound design, requires approved metric contracts for every visual, and never self-grants dashboard_ready: pass. - Fail-closed blockers: Every stop names what is missing, the evidence checked, the responsible owner, and the unblocking action, with no numeric scores. - Use Case: Given an approved Report Intent for a retail subject area, drive gap detection, duplicate-page checks, contract-bound design, report composition, and dashboard QA, then stop with a reviewable blueprint package for the named report owner. ## Quick Start Ask the agent to coordinate the dashboard work end to end for a subject area that already has an approved report intent and a passing semantic model readiness gate.

Frequently Asked Questions about dashboard-intelligence

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build a governed dashboard end to end in Seshat BI?▼

Start with an approved Report Intent and a passing semantic_model_ready gate, then let the coordinator inspect committed state and invoke each shipped capability in order: gap check, duplicate planning, design, composition, and QA. It stops at the human blueprint review.

What happens if the semantic model readiness gate is not passing?▼

The coordinator treats semantic_model_ready values of warning, blocked, or not_started as a hard failure and stops. It names the missing readiness, the evidence checked, the responsible owner, and the unblocking action instead of designing any data-bound visual.

Can the agent approve the dashboard blueprint itself?▼

No. The dashboard_blueprint_approval decision is a named-human action recorded in the Decision Store, and an agent identity never satisfies the approved_by requirement. The highest automated outcome is stopping at the human blueprint review seam.

Does this skill connect to Power BI or a live database?▼

No. The skill is read-only and static: it opens no database or Power BI Desktop connection, authors no PBIR, and performs no publish, refresh, or export. Execution is a deferred boundary; the journey ends at committed on-disk artifacts.

What is an orphan visual and why does it block progress?▼

An orphan visual is a visual with no approved metric contract behind it. The coordinator stops on any orphan because the quality criteria require zero orphan visuals; the visual must be bound to an approved contract or dropped before design continues.