AgentSkillsCN

Research Synthesis

研究综述

SKILL.md

Research Synthesis

Synthesize qualitative and quantitative research into structured insights. Analyze interview notes, survey responses, support tickets, or any unstructured data to identify themes and actionable findings.

Usage

Share research data (interview notes, survey results, feedback, support tickets) and I will extract themes, identify patterns, and produce structured insights.

Capabilities

  • •Thematic analysis of qualitative data (interviews, feedback, notes)
  • •Survey data interpretation (quantitative and open-ended)
  • •Cross-source triangulation for stronger findings
  • •Persona development from behavioral patterns
  • •Opportunity sizing and prioritization

Thematic Analysis Method

  1. •Familiarize: Read through all data to get the overall landscape
  2. •Code: Tag each observation, quote, or data point with descriptive codes
  3. •Develop themes: Group related codes into candidate themes
  4. •Review: Check themes against data -- sufficient evidence? Distinct from each other?
  5. •Refine: Define and name each theme clearly with 1-2 sentence description
  6. •Report: Write up themes as findings with supporting evidence

Extracting Insights from Notes

For each source, identify:

  • •Observations: What was described, experienced, or felt? Note context (when, where, how often)
  • •Direct quotes: Verbatim statements that powerfully illustrate a point
  • •Behaviors vs. stated preferences: What people DO often differs from what they SAY
  • •Signals of intensity: Emotional language, frequency of issue, workarounds, impact when things go wrong

Cross-Source Analysis

  • •Look for patterns that appear across multiple sources
  • •Note frequency: how many sources mention each theme
  • •Identify segments: do different groups show different patterns
  • •Surface contradictions: disagreements often reveal meaningful segments
  • •Find surprises: what challenges prior assumptions

Triangulation

Strengthen findings by combining:

  • •Method triangulation: Same question, different methods (interviews + survey + analytics)
  • •Source triangulation: Same method, different participants
  • •Temporal triangulation: Same observation at different points in time

A finding supported by multiple sources is much stronger than one from a single source.

Presenting Findings

Insight Format

code
**Theme: [Name]**
[1-2 sentence description of what this theme captures]

Evidence:
- [Observation 1 with source attribution]
- [Observation 2 with source attribution]
- [Supporting quote]

Frequency: Mentioned by N of M sources
Confidence: High/Medium/Low
Implication: [What this means for decisions]

When Sources Disagree

Report the disagreement honestly. Check if it is due to different populations, stated vs. actual preferences, or measurement differences. Investigate further rather than silently picking one version.

Example Prompts

  • •"Analyze these interview notes and find the key themes"
  • •"Synthesize this customer feedback into actionable insights"
  • •"What patterns do you see in these support tickets?"
  • •"Combine these survey results with the interview findings"
  • •"Build user personas from this research data"

Tools

  • •read_file: Read research data, notes, and survey results
  • •write_file: Save synthesis reports and findings
  • •edit_file: Update existing research documents
  • •list_dir: Find research files in a directory
  • •web_search: Research industry benchmarks or related findings