research-driven-development

To manage and document research, you can use a structured six-step process that scopes, decomposes, synthesizes, writes, and closes research tasks. This ensures all findings are preserved as persistent documentation in your docs directory.

Synthesizing knowledge from parallel research agents involves dispatching sub-agents for broad and focused inquiries, then aggregating their findings. The process generates persistent documents with summaries, key findings, comparisons, and actionable recommendations.

Generating persistent documentation for AI research tasks requires writing the synthesized findings into a dedicated docs/research/ directory. This creates a permanent record with summaries, key findings, and recommendations for future reference.

Yes, you can use parallel research agents to decompose broad knowledge inquiries. The system triggers on research or planning prompts, dispatching focused sub-agents to ensure comprehensive coverage before synthesizing the results into a unified document.

Yes, research-driven development operates without external dependencies. It uses internal scripts, references, and assets to manage the entire six-step research workflow, from initial scoping to final document generation, entirely within your local environment.

You should avoid using a structured research process for simple, single-query lookups that do not require persistent documentation. The overhead of dispatching parallel agents and following a six-step synthesis workflow is designed for complex, multi-faceted research tasks.