search-first

The `search-first` skill streamlines the workflow of researching existing solutions before writing custom code. It addresses the common problem of developers reinventing the wheel by ensuring that known libraries, tools, and patterns are evaluated before new implementations are created. By systematizing the search and evaluation process, this skill helps maintain code quality, reduces development time, and ensures that solutions are aligned with best practices and community standards.

This skill offers a structured five-step workflow: need analysis, parallel search, evaluation, decision-making, and implementation. It leverages a researcher agent to perform searches across npm, PyPI, GitHub, and internal MCP/skill repositories. The skill includes a decision matrix to guide whether to adopt, extend, compose, or build custom solutions based on factors such as functionality, maintenance, documentation, community support, and licensing. Users can employ a quick mental checklist for common scenarios or invoke a full agent mode for more complex requirements, producing structured comparisons and recommendations. It also provides category-specific search shortcuts for development tooling, AI/LLM integration, data and APIs, and content processing.

Typical use cases include starting new features with potential existing solutions, adding dependencies or integrations, or creating new utilities and helpers. The target users are developers, technical leads, and software engineers who want to ensure efficient and informed coding practices. Additionally, integration with planner agents allows for informed architecture planning by incorporating available tools early in the development process, avoiding redundant implementations and improving project efficiency.