Skip to content
Get packs

Guide · Agent skill

How to use the AI Studio Image agent skill: setup and examples

Updated · written with AI from the project's own documentation and checked against our catalogue data

The AI Studio Image skill is a packaged set of instructions and Python scripts that lets an agent such as Claude generate photos with Google AI Studio (Gemini and Imagen models). It is aimed at social media creators who want lifestyle or influencer-style images, educators who need clean visuals for course material, and developers who already have a Google AI Studio API key. This guide explains what the skill does, how to install it, how to ask for images, what it needs, and where the risks are.

AI Studio Imagesickn33/agentic-awesome-skills · MIT

Quick start

  1. 1Get an API key. Create or copy a Google AI Studio API key at https://aistudio.google.com/apikey.
  2. 2Install the skill. Unzip the downloaded ZIP into ~/.claude/skills/ or a project's .claude/skills/ folder, or upload it under Settings → Capabilities → Skills.
  3. 3Set GEMINI_API_KEY. Store the key as an environment variable or in a .env file in the skill root.
  4. 4Install dependencies. Run pip install -r scripts/requirements.txt with Python 3.10 or newer.
  5. 5Ask for an image. Make a plain request using a trigger phrase such as "gera imagem" or "foto influencer".
  6. 6Review and iterate. Check the result and ask for changes to lighting, framing, realism or setting.

What is an agent skill and what does this one do?

An agent skill is a folder with a SKILL.md file and optional scripts that an AI agent loads when a request matches the skill's description. The agent then follows the instructions in that file instead of improvising.

This skill, ai-studio-image, handles image requests. When you ask for a picture, the agent first rewrites your plain request into a longer prompt meant to make the result look like an unedited smartphone photo. It then runs bundled Python scripts that call Google's image models and save the output to a local folder.

  • Prompt humanization layers: device and technique, natural lighting, human imperfections, authentic skin and expressions, and real environmental context.
  • Two modes: influencer (social posts, lifestyle, personal branding) and educational (course material, tutorials, slides, infographics).
  • Output formats: square (1:1), portrait (3:4), landscape (16:9) and stories (9:16).
  • Preset templates such as cafe-lifestyle, mirror-selfie and food-flat-lay, listed by templates.py --list.

How do you install the ai-studio-image skill?

Download the ZIP for this skill from the directory page and unzip it. Then place the ai-studio-image folder in one of these locations: ~/.claude/skills/ for all your projects, or .claude/skills/ inside a specific project for that project only.

If you use the Claude apps instead, open Settings, go to Capabilities, then Skills, and upload the skill there.

  • Personal install: unzip into ~/.claude/skills/
  • Project install: unzip into .claude/skills/ in your project folder
  • Claude apps: Settings → Capabilities → Skills → upload

How do you set up the Google AI Studio API key and dependencies?

Get a key at https://aistudio.google.com/apikey: click Create API Key, select or create a Google Cloud project, and copy the key. Store it as the GEMINI_API_KEY environment variable, or in a .env file in the skill's root folder.

Install the Python dependencies listed in scripts/requirements.txt. The source lists google-genai, Pillow and python-dotenv, and says Python 3.10 or newer is required.

  • Python 3.10+ is required by scripts/requirements.txt.
  • Key location: GEMINI_API_KEY as an environment variable, or a .env file in the skill root.
  • Quick test: run the generate.py script with --list-models to confirm the setup works.
pip install -r <skill-root>/scripts\requirements.txt

How do you trigger the skill with natural requests?

You do not need to type commands. The skill is designed to activate when your request contains phrases such as "gera imagem", "gerar foto", "criar imagem", "foto realista", "imagem humanizada" or "foto influencer". These are Portuguese trigger phrases from the original skill, so requests in Portuguese are the most reliable way to activate it.

The agent will usually ask which mode and format you want, or infer them from context. A request that mentions stories suggests the 9:16 format, and a feed post suggests 1:1.

  • Example: "gera imagem de uma mulher tomando café numa cafeteria, formato stories"
  • Example: "criar imagem educacional de um diagrama de ciclo da água para slides"
  • Example: "foto influencer de um workspace minimalista, feed"

How do the models, costs and paid flag work?

The skill lists six models. Only gemini-2-flash-exp is listed without the paid flag, and the source describes its cost as "to be verified" and notes it is a historical snapshot whose availability you should check. The other five require the --force-paid flag and are priced per image in the source: imagen-4 at $0.03, imagen-4-ultra at $0.06, imagen-4-fast at $0.02, gemini-flash-image at $0.039 and gemini-pro-image at $0.134.

Check your billing setup before you run paid models. The skill's requirements state that Imagen and pro image models need paid billing.

  • gemini-2-flash-exp: no paid flag listed; cost not confirmed in the source
  • imagen-4-fast: $0.02 per image, for high volume
  • imagen-4: $0.03 per image, high quality
  • gemini-flash-image: $0.039 per image, for image editing
  • imagen-4-ultra: $0.06 per image, for print and 2K
  • gemini-pro-image: $0.134 per image, for 4K reference work
python <skill-root>/scripts\generate.py --prompt "prompt humanizado" --mode influencer --format square --model gemini-2-flash-exp --output <skill-root>/data\outputs\

What safety and privacy points should you know?

This skill runs scripts and uses the network. The scripts send your prompts to Google's services using your API key and save images to a local folder. Read the scripts before running them, and treat the API key as a secret. The source also includes a rule that the agent should not run a step until its required files exist.

The skill is built to make AI output look like an unedited phone photo. Do not present these images as real photographs of real people, and do not use them to mislead others. Check the rules of any platform where you publish the images, and disclose AI generation where your audience or context expects it.

  • Runs scripts: yes. Uses network: yes.
  • Keep GEMINI_API_KEY out of shared files and public repositories.
  • The catalog copy does not include references/api-reference.md. Steps that depend on it are unavailable until you obtain the original file from a verified source.
  • The skill's risk rating in its metadata is critical; review its files before installing.

What tips help you get better results?

Describe the scene in a few plain sentences rather than a list of keywords. The source's prompt guide quotes Google's principle: "Describe the scene, don't just list keywords." Name the setting, the light, the camera and the subject's activity.

Iterate with specific adjustments. The skill's workflow offers four kinds of change: relight, reframe, adjust how natural the image looks, or change the setting.

  • Start from a template with templates.py --list, then adjust the prompt.
  • Ask for changes one at a time: lighting, framing, realism level or setting.
  • Match the time of day to the activity, for example golden hour for a portrait.
  • Use a fixed output folder so you can find your images later.
python <skill-root>/scripts\templates.py --list

What people use it for

  • Lifestyle posts for a personal brand, such as a person in a cafe with a drink.
  • Clean visuals for course slides or a tutorial, using educational mode.
  • Stories or Reels visuals in 9:16 format for a social channel.
  • A scripted image pipeline for a developer who already has a Google AI Studio key.

Frequently asked questions

How do I install the ai-studio-image skill?

Download the ZIP from the directory page and unzip it into ~/.claude/skills/ for personal use or into .claude/skills/ in a project. In the Claude apps, go to Settings → Capabilities → Skills and upload it there.

Is the ai-studio-image skill free?

The skill is listed under the MIT licence. Image generation is not free in every case: the source lists per-image prices for five models that require the --force-paid flag, and the cost of gemini-2-flash-exp is marked as to be verified.

What do I need to run it?

You need Python 3, the packages in scripts/requirements.txt (google-genai, Pillow and python-dotenv), a GEMINI_API_KEY from Google AI Studio, and paid billing for Imagen and pro image models.

Where do I get a GEMINI_API_KEY?

Go to https://aistudio.google.com/apikey, click Create API Key, select or create a Google Cloud project, and copy the key. Store it in an environment variable or in a .env file in the skill root.

Why does the skill need network access?

The scripts send your prompt to Google's Gemini or Imagen image services and receive the generated image back. The source states that the skill runs scripts and uses the network.

What should I do if a step fails because a file is missing?

The catalog copy does not include references/api-reference.md. The source says not to run a dependent step until its required file exists, and to obtain the original supporting material from a verified source rather than creating substitute files.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

Use this skill if you create social or educational images, already have a Google AI Studio key and are comfortable running Python scripts. Pick something else if you cannot pay for image models, need to present images as real photographs, or cannot obtain the missing api-reference.md file that some workflows depend on.

Download the skill