By 2901were
Gemini Creative — an MCP-driven image generation toolkit with auto-activating skills for game asset workflows, UI mockup series, character sprites, and prompt engineering. Wraps Google's Gemini 3 image models (Nano Banana 2 default, Nano Banana Pro) behind a sessions-first API designed for LLM agent consumers.
This skill should be used when the user asks to "design a game character", "create character sprites", "build a character series", "design character variations", "maintain character consistency across poses", "generate a character pose chain", or otherwise needs a full character-design workflow (style bible → base pose → variations) for game development. Covers cross-session consistency, palette locking, and pose chains.
This skill should be used when the user asks to "configure the Gemini API key", "set up GEMINI_API_KEY", "troubleshoot Gemini authentication", "fix MCP not configured errors", "check Gemini API key status", "the API key isn't working", or otherwise needs API-key setup or auth troubleshooting specifically for the gemini-creative-plugin. Covers the three configuration methods (env var, config file, configure_gemini_token tool), priority order, and diagnostic checks via get_configuration_status.
This skill should be used when the user asks to "generate game assets", "create pixel art", "make a sprite", "design a tileset", "generate a tier evolution atlas", "create a merge-game asset pack", "design game items", or otherwise needs to produce pixel art / game art via the gemini-creative-plugin. Covers the Identity Guide pattern (with pixel-art-specific alternative), tier-atlas generation, isolated-subject extraction patterns, and guide overreach workarounds. Palette/resolution tables and sprite/tileset workflows live in references/.
This skill should be used when the user asks to "write a prompt for image generation", "design a generation prompt", "create a sprite prompt", "compose a Gemini prompt", "structure a prompt for game art", or otherwise needs to construct prompts for the gemini-creative-plugin. Covers the Identity Guide 5-line pattern, isolation suffix template for asset extraction, narrative-vs-keyword rules, and Do/Avoid prompting rules. Domain-specific prompt structures (sprites, environments, UI, icons, web) live in references/domain-structures.md.
This skill should be used when the user asks to "generate one image", "make a single image", "create a quick image", "generate an image quickly", "just need one image", or otherwise needs a single-shot image generation without session setup. Covers the minimal generate_image flow and when to skip the session machinery.
Admin access level
Server config contains admin-level keywords
Requires secrets
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Needs API keys or credentials to function
Needs API keys or credentials to function
A Claude Code plugin wrapping a Model Context Protocol (MCP) server for AI image generation using Google's Gemini 3 image models. Designed for LLM agents — built around sessions, organized file output, and 9 auto-activating skills covering game art, UI mockups, character sprites, and prompt engineering.
Version: v0.5.1 · Default model: Gemini 3.1 Flash (Nano Banana 2) · Also supports: Gemini 3 Pro · License: MIT
Two slash commands inside Claude Code:
/plugin marketplace add https://github.com/2901were/gemini-creative-plugin
/plugin install gemini-creative-plugin@gemini-creative-marketplace
Then set your Gemini API key (get one at aistudio.google.com/apikey):
export GEMINI_API_KEY="your-key-here"
That's it. The plugin registers the MCP server and the 9 auto-activating skills. Restart Claude Code after setting the env var (MCPs inherit env at launch).
Image generation:
generate_image — text-to-image with optional reference images (max 14, identity/style anchors), aspect ratio, resolution, and thinking-level controlsedit_image — modify an existing image with text + up to 14 reference imagescontinue_editing — iterate on the last image without re-specifying its pathSessions:
start_creative_session / send_creative_message / end_creative_session — organize multi-image workflows with shared defaults (model, aspect ratio, output directory)get_session_info — list active sessions or inspect oneSetup + utilities:
configure_gemini_token / get_configuration_status — API-key setupget_last_image_info — file path, size, and timestamp of the last generated imageget_prompt_template — pre-built prompt structures for common asset typesClaude Code triggers these automatically when the conversation matches their description:
| Skill | Triggers on |
|---|---|
/gemini-quick | Single image, fast one-off |
/gemini-characters | Game characters — style bible, poses, cross-session consistency |
/gemini-sprite-series | Consistent character sprites across poses or animation frames |
/gemini-game-assets | Pixel art, tilesets, palettes, isometric, tier evolutions |
/gemini-ui-mockups | Multi-screen UI mockups with style selection |
/gemini-session | Session setup, model selection, troubleshooting |
/gemini-workflows | Workflow decision tree across the four main patterns |
/gemini-prompts | Domain-specific prompt structures (sprites, UI, icons, web) |
/gemini-configure | API-key setup and authentication troubleshooting |
For visual consistency across multiple images, always pass the previous image via the images: [path] parameter in send_creative_message. Without it, Gemini reconstructs from text alone and details drift between calls — proportions shift, colors change, fine features vary.
Prompt structure:
[Style] [Subject] in [pose/state], [composition], [distinctive details], [background], [technical specs]
Pixel-art bible prefix:
2D pixel art, 32x32 sprite, PICO-8 16-color palette, 1px black outline, 2-tone cel shading, transparent background —
"Gemini API token not configured" — Run get_configuration_status to verify setup. Set GEMINI_API_KEY env var and restart Claude Code (MCPs inherit env at launch).
Images generating inconsistently across multiple calls — Pass the previous image via the images: [path] parameter in send_creative_message. The guide-pinning approach + image references are what locks consistency.
"Maximum sessions (5) reached" — Run get_session_info and end an older session.
Array parameters not working — Ensure your MCP client sends arrays as JSON arrays, not strings. The server handles both formats defensively, but the JSON form is canonical.
Built by Vladimir Bulanenko. MIT licensed.
npx claudepluginhub 2901were/gemini-creative-plugin --plugin gemini-creative-pluginOpenAI Creative — an MCP-driven image generation toolkit with auto-activating skills for game asset workflows, UI mockup series, character sprites, and prompt engineering. Wraps OpenAI GPT Image models behind a sessions-first API designed for LLM agent consumers.
Reliable automation, in-depth debugging, and performance analysis in Chrome using Chrome DevTools and Puppeteer
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Access thousands of AI prompts and skills directly in your AI coding assistant. Search prompts, discover skills, save your own, and improve prompts with AI.
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
Intelligent draw.io diagramming plugin with AI-powered diagram generation, multi-platform embedding (GitHub, Confluence, Azure DevOps, Notion, Teams, Harness), conditional formatting, live data binding, and MCP server integration for programmatic diagram creation and management.