How to use the Infographics agent skill: setup and examples

What an agent skill is and what this one does

An agent skill is a folder of instructions and scripts that an AI agent can load when a task matches it. This one, from K-Dense-AI/scientific-agent-skills, is named infographics and is licensed MIT.

Once installed, the agent can turn a topic or your supplied data into an infographic image. It sends a description to an image model (Nano Banana 2 via OpenRouter), has a separate AI model (Gemini 3.7 Flash) review the result against a quality threshold for the document type, and regenerates if the review requests improvements, within an iteration limit.

• Supports ten layouts: statistical, timeline, process, comparison, list, geographic, hierarchical, anatomical, resume and social.
• Supports eight style presets, reference images and colorblind-safe palette starting points.
• Optionally gathers candidate facts first with Perplexity Sonar Pro. The skill preserves the source records but does not independently verify them.

Install the infographics skill

Download the ZIP from the skill page and unzip it. For a personal install, place the skill folder in ~/.claude/skills/. For a single project, place it in that project's .claude/skills/ folder.

In the Claude apps you can instead upload the skill under Settings → Capabilities → Skills.

• Python 3.10 or newer is required.
• Install the one runtime dependency with: python -m pip install requests
• Set the OPENROUTER_API_KEY environment variable. Generation, review and research all use paid OpenRouter API calls.

How to trigger the skill with natural requests

You do not need to type script commands. Describe the infographic you want and mention the kind of content, such as statistics, a timeline or a process. The agent loads the skill when the request matches its description.

If accuracy matters, say so and ask for research. Be explicit about the numbers you already have, because supplied data is usually more reliable than gathered facts.

• "Make a statistical infographic of heart disease rates worldwide, using a colorblind-safe palette."
• "Create a timeline infographic of our company milestones from 2010 to 2025 in corporate style."
• "Draft a marketing comparison graphic for these two products. Use a high review standard."

Run it from the command line

The scripts can also be run directly. Output must be a PNG path. The examples below are illustrative; paid generation was not executed during the skill's review, so check output before relying on it.

The --doc-type option sets the review threshold: marketing 8.5/10, report 8.0/10, presentation 7.5/10, social and internal 7.0/10, draft 6.5/10, and default 7.5/10.

Add --research to gather candidate facts with Perplexity Sonar Pro before generating.

• Output is versioned PNG files plus a review log showing models, scores, quality_met and termination_reason.
• A saved image can still be unreviewed or below threshold, so always check it.

What the skill needs and safety notes

The skill runs Python scripts and uses the network. Generated content and prompts are sent to third-party AI services through OpenRouter, and research requests go to Perplexity Sonar Pro. Do not use it with content you are not allowed to send to those services, and do not use it where paid API usage is not acceptable.

Review the scripts before running them in a sensitive environment. The skill's files include scripts/generate_infographic.py and scripts/generate_infographic_ai.py, plus reference documents in references/.

• Requires Python 3.10+, the requests package, network access and OPENROUTER_API_KEY.
• Scripts run locally and make network calls to OpenRouter.
• Keep your API key out of shared files and logs.

Tips for better results and checking output

Supply exact numbers, units, denominators and dates in your request, and replace sample figures with verified, dated sources. Generated images can misstate numbers or text, so inspect every final image at the size it will be shown.

For technical diagrams such as circuits, biological pathways or neural network architectures, use the scientific-schematics skill instead. For exact geometry in scientific figures, supply reference images.

• Choose a document type that matches the audience so the review threshold fits.
• Use a colorblind-safe palette such as wong for accessibility.
• For scientific or medical publication, verify primary sources before using any number.