research

Academic research specialist for scholarly sources, peer-reviewed papers, and scientific literature. Use when research requires academic rigor — literature reviews, citation analysis, methodology evaluation, or understanding the state of scholarly consensus on a topic. <example> Context: User needs to understand the current research landscape on a scientific topic user: "What does the research say about transformer architectures for time series forecasting?" assistant: "I'll use the academic-researcher agent to survey the peer-reviewed literature and identify the key findings and open questions." <commentary> Understanding scholarly consensus requires systematic literature search and quality assessment. </commentary> </example> <example> Context: User needs to evaluate the strength of evidence for a claim user: "Is there solid evidence that retrieval-augmented generation reduces hallucination in LLMs?" assistant: "I'll use the academic-researcher agent to find and evaluate the peer-reviewed studies on RAG's effect on factual accuracy." <commentary> Evidence evaluation requires tracking citations, assessing methodology, and distinguishing strong from weak findings. </commentary> </example> <example> Context: User needs a literature review for a proposal or paper user: "I need a literature review on differential privacy in federated learning" assistant: "I'll use the academic-researcher agent to conduct a systematic review of the academic literature on this intersection." <commentary> Literature reviews require academic database searching, citation network analysis, and gap identification. </commentary> </example>

Deep multi-source research agent for thorough investigation of any topic. Use when a question requires cross-referencing multiple source types (academic, industry, news, government), structured analysis, and a well-cited report with explicit confidence assessments. <example> Context: User needs a thorough briefing on an unfamiliar topic user: "Give me a comprehensive briefing on the current state of nuclear fusion energy" assistant: "I'll use the comprehensive-researcher agent to investigate across academic, industry, government, and news sources and produce a structured report." <commentary> Broad topics requiring multiple source types and structured synthesis are this agent's core purpose. </commentary> </example> <example> Context: User needs to make an informed decision about a complex topic user: "What are the real trade-offs of adopting WebAssembly for our backend services?" assistant: "I'll use the comprehensive-researcher agent to research technical, business, and ecosystem perspectives and present the evidence." <commentary> Decision support requiring balanced multi-perspective analysis benefits from structured research methodology. </commentary> </example> <example> Context: User needs due diligence on a technology or trend user: "Research the current state of AI code generation tools and their impact on developer productivity" assistant: "I'll use the comprehensive-researcher agent to survey academic studies, industry reports, and practitioner experience across this space." <commentary> Due diligence across a broad space requires systematic decomposition and multi-source triangulation. </commentary> </example>

Report generation agent that transforms research findings into polished, well-structured documents. Use as the final step after research and synthesis are complete — takes raw findings and produces a readable report with proper citations, clear narrative flow, and audience-appropriate language. <example> Context: Research is complete and needs to be turned into a deliverable user: "Turn these research findings into a report for our engineering leadership" assistant: "I'll use the report-generator agent to create a structured report tailored for an engineering leadership audience." <commentary> Transforming raw findings into audience-appropriate deliverables with proper structure and citations. </commentary> </example> <example> Context: User needs findings formatted as a specific report type user: "Create a technical comparison report from this library evaluation data" assistant: "I'll use the report-generator agent to produce a comparison report with evaluation tables and trade-off analysis." <commentary> Different report types (technical, comparison, executive briefing) require different structures and emphasis. </commentary> </example> <example> Context: Synthesized research needs a final polished form user: "Package up all our AI safety research into a comprehensive report with proper citations" assistant: "I'll use the report-generator agent to create the final document with executive summary, structured findings, and full bibliography." <commentary> The final step in a research pipeline — converting synthesis into a polished deliverable. </commentary> </example>

Research synthesis agent that consolidates findings from multiple sources or parallel research agents into a unified analysis. Use after dispatching multiple researcher agents in parallel to merge their findings into a coherent picture that preserves all perspectives. <example> Context: Multiple research agents returned findings on the same topic user: "I dispatched academic, technical, and market researchers on quantum computing — synthesize their findings" assistant: "I'll use the research-synthesizer agent to consolidate all three research streams into a unified analysis." <commentary> Multi-agent research produces separate findings that need to be merged, deduplicated, and cross-referenced. </commentary> </example> <example> Context: User has gathered information from diverse sources and needs it unified user: "I have notes from 8 different sources on this topic — help me make sense of it all" assistant: "I'll use the research-synthesizer agent to identify themes, contradictions, and evidence strength across your sources." <commentary> Synthesis requires identifying patterns across sources while preserving important disagreements and nuance. </commentary> </example> <example> Context: Research has been conducted and the user needs the "so what" user: "We've done the research — now tell me what it all means together" assistant: "I'll use the research-synthesizer agent to extract the cross-cutting themes, highlight where sources agree and disagree, and assess the overall evidence quality." <commentary> Moving from individual findings to integrated understanding is the synthesis agent's core function. </commentary> </example>

Technical research agent for analyzing code repositories, libraries, APIs, and implementation approaches. Use when evaluating open source projects, comparing technical solutions, assessing code quality, or researching implementation patterns and best practices. <example> Context: User needs to choose between competing libraries user: "I need to pick a Go HTTP router — compare chi, gorilla/mux, and the standard library" assistant: "I'll use the technical-researcher agent to analyze each project's architecture, maintenance status, performance characteristics, and community adoption." <commentary> Library comparison requires examining repos, benchmarks, API design, and community health indicators. </commentary> </example> <example> Context: User wants to understand how a specific technology works user: "How does SQLite's WAL mode actually work under the hood?" assistant: "I'll use the technical-researcher agent to research the implementation details from the source code and documentation." <commentary> Understanding implementation internals requires analyzing source code, technical docs, and expert discussions. </commentary> </example> <example> Context: User needs to evaluate whether a project is production-ready user: "Is Drizzle ORM mature enough for production use?" assistant: "I'll use the technical-researcher agent to assess its stability, test coverage, maintenance cadence, and community adoption." <commentary> Production readiness assessment requires evaluating code quality, issue tracker health, release cadence, and real-world adoption. </commentary> </example>