agentsclimarketplace

Frappe insights

Skill iam-nazmul/erpnext-toolkit/skills/frappe-insights

Set up and use Frappe Insights — the open-source BI/analytics app for Frappe/ERPNext — for data sources, query building (visual or SQL), charts, and dashboards. Use this skill whenever the user mentions Insights, wants BI/analytics/dashboards over ERPNext data, needs to connect an external MariaDB/PostgreSQL database or upload CSVs for analysis, asks about the Insights query builder, calculated columns/expressions, joins across DocTypes, sharing or embedding dashboards, or Insights permissions/teams. Also use when deciding between Insights and built-in ERPNext reports (Query/Script Reports, Desk dashboards) — Insights is for exploratory/self-serve analytics; see the erpnext skill's reports.md for the built-in options.From its SKILL.md

Install
npx -y skills add iam-nazmul/erpnext-toolkit --skill frappe-insights

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

3.4 KB, as published. Nobody here has run it

Frappe Insights (BI & dashboards)

Insights (github.com/frappe/insights) is a Frappe app that adds a self-serve BI tool to a bench: connect data sources, build queries visually or in SQL, turn them into charts, and compose dashboards — all against the site's own database and/or external databases. Its frontend is the same Vue 3 + frappe-ui stack; its data lives in Insights DocTypes on the site.

When Insights vs built-in reporting

NeedUse
Operational list with filters, printableERPNext Report Builder / Query Report (erpnext skill, reports.md)
Fixed KPI cards on a Desk workspaceDesk Dashboard / Number Card
Ad-hoc exploration, drag-drop charts, joins across tables, non-developer analystsInsights
Analytics over an external database or uploaded CSVsInsights (only option in the Frappe stack)

How to use this skill

TaskReference file
Install/update, first-run setup, connecting data sources (site DB, MariaDB, PostgreSQL, CSV), permissions to grantreferences/setup-and-data-sources.md
Building queries — visual builder, joins, filters, aggregation, calculated columns/expressions, raw SQL — and chartsreferences/queries-and-charts.md
Dashboards, dashboard filters, sharing (private/public/embed), teams & access control, scheduled alertsreferences/dashboards-and-sharing.md

Ground rules

  • Version matters a lot. Insights v3 (current) reorganized the UI around workbooks (queries + charts + dashboards in one document) and runs query processing through a local DuckDB-backed engine; v2 had standalone Query/Chart/Dashboard documents. Concepts carry over but names and screens differ — check the installed version (bench version / app's __init__.py) before giving click-path instructions.
  • Read access is not optional. Queries run with the data source's credentials, not the viewing user's DocType permissions. Anyone who can query a data source can read whatever its DB user can read — scope external DB users to read-only and only the schemas needed, and treat "who can use this data source" as the real permission boundary.
  • Insights queries hit the database directly (SQL), bypassing Frappe's ORM — fine for reads, but never point it at a writable credential.
  • For heavy queries on a production ERPNext site, prefer a read replica or off-peak scheduled refreshes; a runaway analytical query can starve the site's DB.
  • The site DB is MariaDB: DocType tables are tab<DocType> (e.g. tabSales Invoice), child tables link via parent/parenttype/parentfield — the same schema knowledge as Query Reports applies.

Keep looking

Skills are one crate of 609,689. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.