api-pagination_skill

Overview

This skill implements efficient API pagination using offset, cursor, and keyset strategies for large datasets. It provides production-ready patterns and response formats to support paginated endpoints, infinite scroll, and performance-sensitive collection queries. The implementation is TypeScript-friendly and designed for integration with modern stacks like Cloudflare, React, and Tailwind-based frontends.

How this skill works

The skill supplies three pagination strategies: offset/limit for simple pages, cursor pagination for infinite scroll and real-time feeds, and keyset pagination for high-performance access to large collections. Each strategy includes request parsing, safe limits, and a consistent response shape with data, pagination metadata, and navigation links. Cursor logic encodes/decode opaque cursors; keyset relies on indexed sort keys to avoid OFFSET scans.

When to use it

• Offset/limit when dataset is small and UI requires numbered pages.
• Cursor pagination for infinite scroll or APIs that must provide stable forward traversal.
• Keyset pagination for very large tables where OFFSET is too slow.
• When you need consistent response metadata and navigation links for clients.
• When you must enforce maximum page sizes and protect the database from heavy queries.

Best practices

• Enforce a reasonable max limit (e.g., 100) to prevent heavy responses.
• Prefer cursor or keyset for large datasets to achieve O(1) performance.
• Index the fields used for sorting and keyset lookups.
• Avoid count queries on very large tables; supply approximate totals or defer expensive counts.
• Encode cursors opaquely (e.g., base64) and validate them on input.

Example use cases

• Products listing endpoint with offset pagination and total pages for e-commerce.
• Social feed using cursor pagination to support infinite scrolling in a client app.
• Audit log or events listing using keyset pagination to page through millions of rows quickly.
• API that returns standardized pagination metadata and HATEOAS-style links for navigation.
• Protecting backend resources by capping page size and preventing unbounded requests.