deep-research

Deep research uses a graduated protocol with Level 1 quick checks, Level 2 moderate investigation, and Level 3 parallel deep-dive analysis to systematically resolve complex JUCE challenges that exceed single-source documentation or quick-context resolution.

Multi-agent research is designed for JUCE problems requiring novel investigation beyond forums, context limits, or standard docs—situations where parallel investigation tracks and coordinated findings accelerate resolution beyond what single-source troubleshooting achieves.

Invoke the /research command on your complex JUCE problem to trigger Level 1 quick checks; the system automatically escalates to Level 3 parallel deep-dive investigation when initial checks indicate deeper analysis is needed.

After research findings are identified, hand off results to plugin-improve for applied implementation solutions, bridging investigation insights into concrete code and configuration changes for your JUCE plugin.

Yes; parallel subagents conduct simultaneous investigations across multiple problem domains, then consolidate findings into a unified research output suitable for handoff to implementation workflows.