team-shinchan:research

To gather developer documentation, this skill dispatches a Masumi subagent using the sonnet model to perform web research and return structured output including key findings, linked documentation, and actionable guidance for your queries.

The best way to perform competitive analysis and knowledge-base enrichment is using a research subagent that validates your prompt, delegates execution, and surfaces authoritative references, best practices, and caveats tailored to developer questions.

Web research validation works by checking your submitted prompt for missing queries and prompting you to provide one, while also truncating any requests longer than 2000 characters and issuing a warning before executing the research.

Yes, you can use this research subagent for security recommendations and design pattern lookups, receiving curated findings, linked documentation, and actionable guidance suitable for integration directly into a project knowledge base.

A key limitation when handling web research queries is that the input prompt is restricted to 2000 characters; requests exceeding this length are truncated with a warning before the subagent executes the research task.