Setup & Installation
What This Skill Does
Generates read-only MongoDB find queries and aggregation pipelines from natural language requests. Fetches collection schema, indexes, and sample documents through the MongoDB MCP server before building queries, so field names and types are validated against real data.
Reads the actual collection schema and indexes before writing the query, so it catches wrong field names and suggests missing indexes instead of returning silent empty results.
When to use it
- Turning 'users who signed up last week and haven't logged in since' into a find filter
- Building a $group aggregation to sum order totals by category
- Translating a SQL JOIN request into a $lookup pipeline stage
- Writing a geospatial query without mixing up longitude and latitude order
- Converting a date-relative request like 'orders from the past 30 days' into the right ISODate filter