search-first

To find existing libraries before writing new code, perform a parallel search across npm, PyPI, GitHub, and MCP servers to surface tools and patterns, then evaluate candidates using a structured decision matrix to avoid reinventing solutions.

The best way to evaluate whether to adopt, extend, or build a dependency is to apply a structured decision matrix to search results, scoring candidates on integration fit, feature coverage, and guardrails for your JavaScript, Python, or web stack.

Parallel search for existing tools queries npm, PyPI, GitHub repositories, and Claude Skills simultaneously, surfacing libraries and patterns that match your feature requirements to prevent redundant development work.

Yes, you can use this research workflow for Python and JavaScript feature design, as the structured search and evaluation process explicitly supports identifying existing integrations and patterns across JavaScript, Python, and web stacks.

You should build your own solution instead of using existing libraries when the decision matrix indicates that available tools lack required features, integration guardrails are too restrictive, or composition overhead exceeds the effort of a custom build.

To integrate existing GitHub repositories into project planning, use the recommended actions from the evaluation phase to apply research findings, mapping integration points and setting guardrails during feature design and dependency addition.