retail-control-room

Aggregates per-table data-quality evidence into a read-only portfolio control room.

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

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When a retail data warehouse has many tables at different readiness stages, it is hard to see which table is broken, which is blocked, and which one to fix next without opening every per-table status file. This Skill consolidates all committed per-table evidence into one worst-first portfolio view. ## Core Features & Use Cases - Per-table roll-up: One row per table showing stage, status, static WARN counts, live finding counts, open blockers, and next action, sorted worst-first. - Portfolio open-blockers list: Every open blocker across all tables with stage, concrete reason, measured evidence, and named owner, with unassigned blockers flagged. - Full traceability: Every number traces to a committed source file (readiness-status.yaml, data-issues.md, blocking-reasons.md) rather than a recomputation or invented score. - Use Case: Ask "which table do I fix next" and receive a ranked view where live validation errors (V-RC2/V-RC15/V-RC16) outrank static warnings, with each cell traceable to its source path. ## Quick Start Ask the assistant to show the data-quality control room for the repo and list which table to fix next.

Frequently Asked Questions about retail-control-room

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

FAQPage Schema
How do I see data quality status across all tables at once?▼

Invoke the control room to render a worst-first roll-up of every table's stage, status, static WARNs, live findings, open blockers, and next action. It joins existing per-table evidence files rather than running new checks.

How do I decide which table to fix next in a data pipeline?▼

The portfolio open-blockers list ranks blockers worst-first: live validation errors (V-RC2/V-RC15/V-RC16) outrank static warnings. Each blocker shows its stage, concrete reason, measured evidence, and named owner.

Does the control room run validation checks or modify files?▼

No. It is strictly read-only: it runs no new validator, edits no per-table artifact, clears no blocker, and opens no database connection. After a run, git status shows zero modified files.

Can I get a single health or confidence score per table?▼

No. The skill explicitly declines to emit fabricated health scores. It returns the four explicit readiness statuses plus measured counts with their source paths instead, per the no-fake-confidence rule.

What happens when a table has missing or conflicting evidence?▼

Missing evidence is shown as not_started or evidence incomplete, never invented. Conflicts such as a pass status alongside an open blocker are surfaced as findings rather than silently resolved.