search-first

To find existing libraries before coding, use a research-first workflow that runs parallel searches across package registries, GitHub, and academic literature. This helps you discover and adopt available tools instead of writing custom code from scratch.

A decision matrix for software tooling evaluates parallel search results to determine whether you should adopt an existing package, extend a current library, compose multiple tools, or build a custom solution based on your specific project requirements.

Yes, you can search academic literature and npm registries simultaneously. The research workflow utilizes parallel search across multiple channels including package registries and academic databases to consolidate tool discovery.

For complex queries researching code optimization solutions, the best approach is to use a full researcher agent. This mode provides deep search capabilities across GitHub and academic literature compared to quick inline checks.

You should avoid parallel search for tool discovery during quick, trivial implementation tasks. In these scenarios, running a full researcher agent is unnecessary overhead, and a quick inline check of package registries is sufficient.