- Home
- Skills
- Image & Video
- Visual Asset Forge
Works with the AI Agents you already use
Visual Asset Forge
Generate, edit, and batch raster assets through the OpenAI Images API with structured prompt specs, stable taxonomy labels, and validated delivery.
Free
Visual Asset Forge
Example session with this skill installed
We're rebuilding the marketing site for our ceramics brand and I need a full image set. Generate a photorealistic landing-page hero of a matte stoneware mug on a warm oak table with soft morning window light, keeping negative space on the right for headline text. Then take the three product photos in input/ (mug-01.jpg, mug-02.jpg, bowl-01.jpg) and turn them into transparent-background cutouts for the catalog cards — the glaze color and handle shape must stay exact. Finally make a late-sunset lighting variant of the hero. Put everything in output/visual-asset-forge/ and validate before delivery. My OPENAI_API_KEY is exported in the shell.
- Read your context and instructions
- Compiled the visual asset forge
- Generated the image
Summary
Delivered 9 validated raster assets for the site refresh: one hero, three transparent product cutouts, three UI card graphics, and two lighting variants. Each was reviewed against its spec for subject accuracy, text fidelity, and must-avoid rules.
| Asset | Mode | Taxonomy | Before | After |
|---|---|---|---|---|
| hero-main.png | generate | photoreal-natural | stock placeholder | 1536x1024, high quality, right-side negative space |
| cutout-01..03.png | edit | background-remove | solid grey backdrop | transparent PNG, glaze + edges locked |
| card-a/b/c.png | batch | ui-mockup | text-only tiles | consistent palette, 1024x1024 |
| hero-dusk.png | edit | lighting-shift | midday light | late sunset, subject geometry unchanged |
Next steps
- Review cutout edges at 200% zoom — cutout-02 needs one mask pass around the handle
- Approve hero-main or re-run with a single change (lens choice) before ad variants
- Batch spec kept at tmp/visual-asset-forge/jobs.jsonl for the remaining 15 assets
- Final files live in output/visual-asset-forge/ under taxonomy-keyed names
hero-main.png
PNG · 1536×1024
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
The problem
Generating production-ready UI assets and product imagery often results in inconsistent styles, messy file management, and unpredictable output quality. Manual prompting through web interfaces makes it impossible to reproduce results or automate batch asset runs.
What it does
- Generates new raster assets using a structured prompt specification format.
- Edits existing images with strict region control and subject identity locking.
- Automates batch image generation and editing via JSONL job files.
- Handles background removal and transparency workflows with edge preservation.
- Validates deliverables against a stable taxonomy of asset types.
Frameworks & tools
Python, OpenAI Images API, Pillow, and uv or pip for dependency management.
Why this beats prompting it yourself
This skill enforces a formal asset taxonomy and structured intake checklist, ensuring composition, materials, and lighting are defined before the API call. It manages file hygiene between tmp/ and output/ directories and provides a bundled CLI workflow that is easier to audit and script than one-off chat prompts.
Use cases
- Generating consistent hero imagery for landing pages.
- Batch creating product thumbnails for e-commerce catalogs.
- Iteratively replacing backgrounds or localizing text in existing UI mockups.
- Creating transparent product cutouts for design systems.
Known limitations
Does not support vector file editing or purely textual design advice. Requires a valid OPENAI_API_KEY set in the environment.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Image Generation Skills.
Requires your own OPENAI_API_KEY, with image generation billed by OpenAI. In-image lettering and pixel-exact masks sometimes need one or two single-change retries.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
5 uses
Times used to date
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 2 months ago
- Free with an Agensi account
- Accounts
OpenAI
- Software
Python
Permissions
Allowed Hosts
File Scopes
Needs shell, file I/O, and OPENAI_API_KEY access to install SDK dependencies, read source images, call api.openai.com, and write validated assets to project output folders.