search-first

Upfront research before coding identifies existing solutions, patterns, and constraints to reduce code duplication and rework. It is needed when starting new features or integrating external services to ensure architectural decisions are well-informed.

Perform a precode research pass by following a structured workflow: conduct need analysis, execute parallel searches, evaluate trade-offs, make a decision, and proceed to implementation. This surfaces dependencies and patterns before committing to code.

Yes, you can use a research-first workflow for API integrations and evaluating dependencies. It helps evaluate libraries, patterns, patterns, and dependencies before committing to a tech choice, effectively reducing risks and avoiding reinventing the wheel.

The best way to reduce code rework when planning software architecture is implementing a structured research-first workflow. By analyzing needs and evaluating existing solutions before coding, you select the best approach and minimize future duplication.

Avoid a research-first approach for simple, well-understood modifications where implementation paths are already clear. Over-researching trivial tasks introduces unnecessary overhead, whereas the workflow targets complex architectural decisions and external service integrations.