figure-legend-gen

What This Skill Does

The figure-legend-gen skill is a specialized AI-powered utility designed to automate the creation of publication-quality figure legends for scientific research. It leverages vision analysis to interpret complex visual data, including bar charts, line graphs, scatter plots, box plots, heatmaps, and various laboratory imaging formats like microscopy, flow cytometry, and Western blots. By analyzing the structure of your uploaded files, the agent generates structured, academic-grade captions that adhere to standard scientific reporting requirements. This includes specifying the figure number, a concise title, a detailed description of the data shown, the underlying experimental methodology, statistical significance markers (such as P-values), and technical details like scale bars where applicable. The tool ensures consistency in scientific documentation, reducing manual writing time while maintaining the rigorous standards required by high-impact academic journals.

Installation

To integrate this skill into your OpenClaw environment, use the following command in your terminal or command-line interface:

clawhub install openclaw/skills/skills/aipoch-ai/figure-legend-gen

Ensure that you have the necessary dependencies, such as PIL and pytesseract, installed in your working directory to allow the agent to perform local image processing and optical character recognition when necessary.

Use Cases

This tool is primarily designed for researchers, PhD candidates, and data scientists. Common use cases include:

• Drafting supplemental materials for journal submissions where multiple figures need consistent, high-quality descriptions.
• Quickly generating baseline legends for exploratory data analysis charts during lab meetings.
• Standardizing figure formatting across large datasets or multi-panel figures.
• Simplifying the documentation of complex image data, such as confocal microscopy, by automatically identifying the need for scale bar descriptions and experimental parameters.

Example Prompts

1. "OpenClaw, please analyze this image (path: results/fig1.png) and generate a publication-ready legend for a bar chart showing protein expression levels across three conditions."
2. "I've uploaded a scatter plot of our latest patient data. Can you use figure-legend-gen to write a comprehensive legend that includes p-values and describes the correlation between variables? Output it in LaTeX format."
3. "Generate a Chinese-language figure legend for this Western Blot image. Ensure it mentions the antibody concentrations and the primary cell line used in the experiment."

Tips & Limitations

To get the best results, ensure your input images are clear and high-resolution. While the tool is highly capable of identifying chart types and extracting visual features, providing additional context—such as the specific statistical test performed (e.g., ANOVA, T-test) or the experimental conditions—in your prompt will yield significantly more accurate legends. Note that this tool generates text legends only; it does not perform data visualization or file editing. For complex, multi-panel figures, you may need to provide brief textual guidance on the individual panels to ensure the agent maps the description correctly to each subsection.