mcpField
D
data-warehouse-metrics-layer-review

data-warehouse-metrics-layer-review

github.com/sylphxaiData & databases139 installsUpdated today
Repository

About data-warehouse-metrics-layer-review

A skill for teams reviewing warehouse metrics layers and semantic models. It turns KPI, lineage, freshness, and dashboard-trust questions into a structured design with owners, controls, and migration steps.

Key features

  1. 01Metric contract and grain definition
  2. 02Lineage and freshness planning
  3. 03Quality tests and access rules
  4. 04Dashboard publishing and change control
  5. 05Migration and deprecation planning

Use cases

  1. 01Standardizing conflicting dashboard KPIs
  2. 02Reviewing freshness and quality controls
  3. 03Planning a KPI change or backfill

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/sylphxai/skills --skill data-warehouse-metrics-layer-review

After installing, type/data-warehouse-metrics-layer-reviewin Claude Code to confirm it loaded.

Bundled filesagents/openai.yamlreferences/data-warehouse-metrics-layer-patterns.md

Details

From the project README

Data Warehouse Metrics Layer Review

Use this skill to convert warehouse metrics layer, semantic model, KPI definition, lineage, freshness, and dashboard trust questions into a concrete artifact with owners, gates, metrics, and recovery paths.

Workflow

  1. Identify decision questions, source systems, metric consumers, current dashboard conflicts, event tables, warehouse models, BI tools, owners, freshness needs, and compliance constraints.
  2. Read references/data-warehouse-metrics-layer-patterns.md.
  3. Classify metric as financial, product usage, growth funnel, support, reliability, marketplace, experiment, AI cost, or operational health.
  4. Define canonical metric contract, grain, dimensions, filters, owner, lineage, freshness SLA, quality tests, access rules, dashboard publishing, and change log.
  5. Produce metrics-layer design, state machine, decision table, event schema, quality checklist, and migration/deprecation plan.

When not to use

  • Do not use for generic advice the base model already handles without this skill's specific artifact contract.

Guardrails

  • Do not let every dashboard redefine core metrics such as active user, conversion, churn, ARR, support SLA, or usage cost.
  • Do not publish metrics without grain, timezone, identity resolution, exclusions, source lineage, and owner.
  • Do not backfill or redefine a KPI without change log, consumer notification, and historical impact estimate.
  • Do not expose sensitive row-level data through convenience metrics or broad BI access.

Output format

Metrics-layer context:
Audience / source of truth / risk boundary:

Metric contract plan:
| Area | Decision | Evidence | Risk | Owner |
| --- | --- | --- | --- | --- |

Lineage, freshness, and change control:
- <trigger> -> <policy, metric, edge case, support note>

Comments

Markdown supported