search-first

To find existing libraries before writing custom code, use a research-driven development workflow that searches package managers, internal skills, and code repositories in parallel to identify proven solutions. This ensures you adopt or extend existing tools rather than reinventing the wheel.

Evaluating npm packages and GitHub repositories for dependency integration is streamlined by a decision support matrix that compares candidates discovered through parallel searches, helping you choose whether to adopt, extend, or build a custom solution based on research findings.

You need to search for existing solutions when adding new features, integrating dependencies, or creating abstractions that require common utilities like date formatting or API clients, ensuring efficient development by prioritizing proven tools over net-new implementations.

Yes, you can use this research workflow to find an npm package for checking dead links in markdown files by running parallel searches across package managers and code repositories, then evaluating the candidates to integrate a well-maintained library.

Parallel code discovery across package managers and repositories works by simultaneously querying npm, MCP, and GitHub to surface existing libraries and patterns, then guiding you through evaluating candidates to decide between adoption, extension, or custom implementation.

The limitation of a research-first approach to dependency management is that it prioritizes adopting or extending proven tools over net-new implementations, which may not suit highly specialized use cases where existing libraries lack required custom functionality or abstractions.