deep-research

Deep research investigates complex JUCE plugin problems using a graduated protocol that escalates from local checks to parallel multi-agent investigation. It autonomously diagnoses root causes of cryptic errors, build failures, and runtime issues that standard troubleshooting cannot resolve, delivering structured findings with recommended solutions.

Use research when troubleshooting reaches Level 3 difficulty, problems require novel implementation strategies, or issues persist after local investigation. The Skill prevents wasted time on simple issues while ensuring complex or unfamiliar JUCE challenges receive thorough, multi-perspective analysis.

The research Skill spawns parallel subagents to investigate different approaches simultaneously, synthesizing findings into a structured report with pros, cons, implementation complexity, and a recommended strategy. This is ideal for novel algorithmic problems where multiple valid solutions exist.

Graduated research starts with quick local checks (Level 1), escalates to documentation searches (Level 2), then deploys parallel multi-agent investigation (Level 3) for complex problems. Each level provides findings; escalation only occurs if earlier levels cannot resolve the issue.

Problems involving parameter persistence in DAWs, novel DSP implementations, cryptic build errors, complex audio processing workflows, and issues that fail standard troubleshooting benefit from deep research. The Skill handles investigation-heavy scenarios while delegating code changes to implementation tools.