Deep Research Skill

Systematic approach to thorough research on any topic.

Research Process

• Scope: define core questions, boundaries, success criteria, and planned sources.
• Gather: broad overview → targeted deep dives. Tools: WebSearch, WebFetch, Context7, Explore agent. Sources: official docs, academic papers, industry reports, community (GitHub, forums).
• Validate: cross-reference claims across sources, verify publication dates for currency, note conflicts and controversies, assess source credibility and evidence quality (facts vs. opinions, methodology rigor, reasoning).
• Synthesize: executive summary (3-5 bullets) → detailed findings by theme → critical analysis (evidence quality, conflicts, gaps) → practical implications → cite sources with dates → state confidence level.

Special Considerations

For AI/ML, current events, technical evaluations, or business topics: read references/domain-tips.md for domain-specific cautions.

Quality Checklist

Before concluding research, verify:

• Multiple authoritative sources consulted
• Recent information included (check dates)
• Key perspectives represented
• Evidence quality assessed
• Conflicts/controversies noted
• Practical implications identified
• Sources properly cited
• Confidence level stated
• Gaps/limitations acknowledged
• Actionable conclusions provided

Tools to Use

• WebSearch / WebFetch: general information, current events, specific URLs.
• Context7: library/framework documentation.
• Agent (general-purpose subagent) — pinned opus/medium: multi-step investigations synthesize across sources, not just locate, so route them here rather than to Haiku Explore (routing + why → ~/.claude/references/claude-models.md).

Iteration

If conflict, expand search and present multiple viewpoints; if insufficient, try different sources or terms; if outdated, search for recent sources; if gaps, ask the user to clarify.