Daily metric check-in with configurable alerts. Red/yellow/green status for the numbers that matter to you.
Runs a repeatable daily standup against your own metrics definition. Instead of clicking through five dashboards every morning, you get one report in a consistent format: each metric labeled green, yellow, or red according to rules you own.
It’s source-agnostic: point it at BigQuery, PostgreSQL, Snowflake, or wrap a custom script—anything you can express as SQL or a callable that returns numbers the skill can interpret.
Maps raw values to green, yellow, and red using bands you define—e.g., adoption above 40% is green, 20–40% yellow, below 20% red. Best for KPIs with clear healthy ranges.
Tracks progress toward a target or deadline: percent complete, days remaining, or binary “gate” checks. Surfaces when you’re off pace even if the absolute number still looks fine.
Looks across multiple weeks (or days) for direction: improving, flat, or worsening. Catches slow leaks that single-point thresholds miss—usage creeping down 3% a week adds up.
metrics.yamlReads your configuration: queries, thresholds, milestone definitions, trend windows, and display names so the report matches your vocabulary.
Runs SQL (or delegated checks) concurrently via subagents where supported, so a wide dashboard doesn’t serialize into a long wait.
Applies threshold bands, milestone logic, and trend detection to each metric, producing a normalized status plus optional detail strings.
Formats everything into a scannable standup: status, values, WoW or period comparison, alerts, and short suggested actions for reds and yellows.
A formatted report listing each metric with a clear status indicator, the current value, week-over-week (or configured period) comparison, highlighted alerts for anything that needs attention, and concise suggested next steps so you’re not staring at a red number without a hypothesis.
Clone the repo and start using this skill in under 60 seconds.