research-outline

name	research-outline
description	Conduct preliminary research on a topic and generate a structured research outline. First phase of the two-phase deep research workflow.
metadata	{"version":"1.0.0","source":"Weizhena/Deep-Research-skills"}

Research Outline

Conduct preliminary research on a topic and generate a structured research framework with items to investigate and fields to analyze.

Workflow

Step 1: Generate Initial Framework

Based on the research topic, use existing knowledge to generate:

• Main research objects/items list in this domain
• Suggested research field framework (dimensions to analyze)

Present the framework and confirm with the user:

• Need to add/remove items?
• Does field framework meet requirements?

Step 2: Web Search Supplement

Ask for time range (e.g., last 6 months, since 2024, unlimited).

Search the web to:

1. Verify if existing items are missing important objects
2. Supplement items based on missing objects
3. Search for topic-related items within time range
4. Supplement new research dimensions/fields

Step 3: Merge Existing Knowledge

If the user has existing field definitions or prior research, merge them with the generated framework.

Step 4: Generate Research Outline

Produce two structured outputs:

Items List:

• Topic name
• Complete list of research objects
• Execution configuration (batch size, items per batch)

Field Definitions:

• Field categories with descriptions
• Detail level for each field (brief / moderate / detailed)
• Uncertain fields marked for verification

Step 5: Confirm and Save

Present the complete outline for user confirmation before proceeding to deep research phase.

Research Types Supported

• Academic: Paper surveys, benchmark reviews, literature analysis
• Technical: Technology comparison, framework evaluation, tool selection
• Market: Competitor analysis, industry trends, product comparison
• Due Diligence: Company research, investment analysis, risk assessment

Best Practices

• Always verify claims with multiple sources
• Mark uncertain or unverified information clearly
• Structure the outline to enable parallel deep-dive research
• Include source URLs for all supplementary findings
• Design fields that are measurable and comparable across items