research-first

Researching current framework documentation involves spawning a dedicated sub-agent that queries context7 for framework docs and Perplexity for validation signals, returning concise findings to keep your main context clean and actionable for implementation.

The best way to validate framework best practices while keeping your main context clean is to perform heavy research in a separate sub-agent context window, cross-referencing context7 and Perplexity, and returning only concise actionable findings to the main execution flow.

Yes, you can coordinate with context7 for framework documentation and Perplexity for validation signals simultaneously through a dedicated sub-agent that queries both sources in a separate context window and returns concise findings.

You need a dedicated research sub-agent for architectural decisions when you require up-to-date framework guidance and validated approaches before implementation, especially during framework selection, uncertainty-driven design exploration, or API usage pattern validation.

Spawning a research sub-agent does not clutter your main execution context because heavy research is performed in a separate context window, and only concise, actionable findings are returned to keep the main context clean and focused.