DonggangChen

mongodb_usage

2
2
# Install this skill:
npx skills add DonggangChen/antigravity-agentic-skills --skill "mongodb_usage"

Install specific skill from multi-skill repository

# Description

This skill should be used when user asks to "query MongoDB", "show database collections", "get collection schema", "list MongoDB databases", "search records in MongoDB", or "check database indexes".

# SKILL.md


name: mongodb_usage
router_kit: FullStackKit
description: This skill should be used when user asks to "query MongoDB", "show database collections", "get collection schema", "list MongoDB databases", "search records in MongoDB", or "check database indexes".
metadata:
skillport:
category: auto-healed
tags: [aggregation, big data, cleaning, csv, data analysis, data engineering, data science, database, documents, etl pipelines, export, import, json, machine learning basics, migration, mongodb usage, mongoose, nosql, numpy, pandas, python data stack, query optimization, reporting, schema design, sharding, sql, statistics, transformation, visualization]


MongoDB MCP Usage

Use the MongoDB MCP server to integrate database queries into workflows.

Read-Only Access

MongoDB MCP is configured in read-only mode. Only queries and data retrieval are supported. No write, update, or delete operations.

Database Queries

Use mcp__mongodb__* tools for:

  • Listing databases
  • Viewing collection schemas
  • Querying collection data
  • Analyzing indexes

Integration Pattern

  1. List available databases with mcp__mongodb__list_databases
  2. Explore collections with mcp__mongodb__list_collections
  3. Get schema information with mcp__mongodb__get_collection_schema
  4. Query data as needed for analysis
  5. Format results for user consumption

Environment Variables

MongoDB MCP requires:

  • MONGODB_URI - Connection string (mongodb://...)

Configure in shell before using the plugin.

Cost Considerations

  • Minimize database calls when possible
  • Use schema queries before running analysis queries
  • Cache results locally if multiple calls needed
    MongoDB Usage v1.1 - Enhanced

🔄 Workflow

Source: MongoDB Performance Best Practices

Phase 1: Discovery & Inspection

  • [ ] Connection: Verify access with mcp__mongodb__list_databases.
  • [ ] Schema Analysis: Understand data types and structure with mcp__mongodb__get_collection_schema.
  • [ ] Index Check: List existing indexes (with list_indexes or similar query).

Phase 2: Query Construction

  • [ ] Filter: Filter queries over indexed fields (Prefix).
  • [ ] Projection: Select only necessary fields ({ field: 1 }) (Network and RAM saving).
  • [ ] Aggregation: Build $match, $group, $project pipeline for complex analysis.

Phase 3: Performance Check (Explain Plan)

  • [ ] Explain: Check if query performs COLLSCAN (Full scan) or IXSCAN (Index scan).
  • [ ] Optimization: Suggest Compound Index for slow queries.

Checkpoints

Phase Verification
1 Does query respond in under 100ms?
2 Is "In-memory sort" limit exceeded (is there disk usage)?
3 Do Regex queries use the start of the index (anchor ^...)?

# Supported AI Coding Agents

This skill is compatible with the SKILL.md standard and works with all major AI coding agents:

Learn more about the SKILL.md standard and how to use these skills with your preferred AI coding agent.