24 Specialized Research Agents
AI Research Assistant for the Complete Research Lifecycle
"Beyond Modal: Creative alternatives, not obvious choices"
How Agents Work
Agents are automatically triggered by keywords and context, dispatched via the Task tool for specialized research tasks.
Auto-detect research context from your conversation
Parallel execution when tasks are independent
Checkpoint integration ensures human decisions
9 Agent Categories
Organized by research lifecycle stage and function
Foundation
Research questions, theory, paradigm foundations
Literature & Evidence
Literature search, quality appraisal
Study Design
Research design, meta-analysis orchestration
Data Collection
Instruments, interview protocols
Analysis
Statistical, qualitative, mixed methods analysis
Quality & Validation
Humanization verification
Publication & Communication
Publication, humanization, journal matching
Systematic Review
PRISMA automation, RAG pipelines
Cross-Cutting
Research integrity, ethics oversight
Model Routing by Complexity
Agents use appropriate models based on task complexity:
Opus
Complex decisions, theory, orchestration
Sonnet
Standard research tasks, screening
Haiku
Quick validation, RAG building
Agent Invocation
Agents are triggered automatically by keywords or manually via Task tool:
Task( subagent_type="diverga:a1", model="opus", prompt="Refine research question using FINER criteria" )
Parallel Execution Groups
Multiple agents can run simultaneously when tasks are independent:
Research Design
A1 + A2 + A5
Question, theory, paradigm
Literature & Evidence
B1 + B2
Search, appraisal
Publication Prep
G1 + G2 + G5
Journal match, writing, AI audit
SR Pipeline
I0 + I1 + I2 + I3
Orchestration, retrieval, screening, RAG
Explore Agent Categories
Browse detailed documentation for each category:
Ready to Use Specialized Agents?
Start your research with 24 agents covering the complete research lifecycle.