industry-research-with-image

name	industry-research-with-image
description	Guide for creating illustrated industry research reports with AI-generated images. Use when users want to write an industry research report that includes AI-generated images - the workflow involves planning, creating framework, gathering data, writing research content, generating images via subagent, and inserting image references.
license	Complete terms in LICENSE.txt

Industry Research with Image

Create industry research reports with AI-generated images following this workflow:

Workflow

Execute these steps in order:

Step 1: Planning & Framework Creation

1. Use write_todos to plan the report step，you need to clearly define each detailed step, including which tools to use, what tasks to perform, where to save the results, and how to conduct checks.
2. Define the complete report framework including:
   • Introduction/overview
   • Main content sections (each section = one topic)
   • Conclusion
3. Identify image positions: For each section that needs visual illustration, plan:
   • Where the image should be placed (which section)
   • What the image should depict (concept, diagram, timeline, etc.)
   • A descriptive filename for the image (e.g., ai_industry_overview.png)
4. Save the framework to a markdown file (e.g., workspaces/report_framework.md)
5. The framework should look like:

   # [Report Title]
   
   ## 1. Section Title
   ![](./pic/section_image_filename.png)
   Section content...
   
   ## 2. Another Section
   ![](./pic/another_image_filename.png)
   More content...
   

Step 2: Gather Information

1. Understand what user needs
2. Check what tools you have available
3. Use analyze-company skill if needed
4. Use web_search tool to gather data (attention: web_search is expensive, only try once or give up this tool)
5. Use specific tools or skills to gather information for each section

Step 3: Generate Images FIRST (Before Writing Content)

1. Use image-generator-agent subagent to generate all images
2. Critical: Generate images based on the framework, one for each planned position
3. Image requirements:
   • Create illustrations for each planned image position
   • Output to workspaces/pic/ directory
   • Use descriptive filenames matching the framework (e.g., ai_industry_overview.png)
4. Verify all images are saved in pic/ folder

Step 4: Write Report Content

1. Now use writer-agent subagent to write comprehensive markdown content
2. Important: Read the framework file to understand where images should go
3. Write content for each section, placing it after the corresponding image reference
4. The content should flow naturally with the visual elements
5. Topics to cover for each section:
   • Introduction/overview
   • Core concepts and features
   • Advantages and benefits
   • Application scenarios
   • Future outlook
6. Save the draft markdown file (e.g., workspaces/report_draft.md)

Step 5: Insert Image References & Finalize

1. Read the draft markdown file
2. Verify all image references are present in the correct positions
3. Each image reference should look like:

   ## Section Title
   
   ![Image description](./pic/filename.png)
   
   Section content...
   

   !!Path don't include workspaces
4. If any images are missing, generate them now
5. Save as final report file (e.g., workspaces/final_report.md)

Key Points

• Always create framework FIRST - define structure and image positions before writing
• Generate images AFTER framework but BEFORE writing content - this ensures images match the planned structure
• Plan image positions in Step 1: Each major section should have a corresponding image
• Use pic/ subdirectory for images to keep workspace organized
• Edit markdown files using read_file + write_file (not write_todos)
• Image naming: Use descriptive names that match the section (e.g., ai_market_trends.png, ai_technology_stack.png)

Workflow Order (Important)

1. Planning & Framework → 2. Gather Info → 3. Generate Images → 4. Write Content → 5. Finalize
         ↑                         ↑                  ↑                    ↑
    Define structure        Collect data      Create all images    Write around
    + Image positions                                             images

Available Tools

• write_todos: Create and update task lists
• writer-agent: Write comprehensive report content
• image-generator-agent: Generate AI images from descriptions
• read_file: Read local markdown files
• write_file: Save or update markdown files
• ls: List directory contents
• mkdir: Create directories (pic/)
• web_search: Search for information (use sparingly)

Image Generation Tips

When generating images for reports:

1. Match the framework: Generate images for each planned position in Step 1
2. Describe the visual concept clearly
3. Specify style (modern, minimalist, professional, data visualization, etc.)
4. Include key elements that represent the topic
5. Mention color scheme if important
6. Filename convention: Use {topic}_{type}.png format

Example prompt: "A clean professional infographic showing AI industry market growth from 2020-2025, bar chart style with blue gradient, modern minimalist design"

Image Position Planning Example

For an AI industry report:

Framework:
# AI行业研究报告

## 1. 行业概览
![](./pic/ai_industry_overview.png)
行业规模、发展历程...

## 2. 技术架构
![](./pic/ai_technology_architecture.png)
核心技术组成...

## 3. 应用场景
![](./pic/ai_applications_showcase.png)
各行业应用案例...

## 4. 未来趋势
![](./pic/ai_future_trends.png)
技术发展方向...