muapi-nano-banana

muapi-nano-banana is a specialized image-generation skill that applies reasoning-driven, structured creative briefs (in the style of Google's Gemini 3 / Nano Banana Pro) to produce high-fidelity images via the muapi.ai API. It solves the problem of low-quality keyword-stuffed prompts by moving to natural-language, logic-based briefs that describe physics, lighting, and spatial relationships coherently.

The skill centers on a "Perfect Prompt" formula composed of Subject, Action, Context, Composition, Lighting, and Style, each with concrete guidance (for example, specifying camera and lens choices like an 85mm lens at f/1.8, or directional volumetric lighting). Advanced features include negative-constraint logic phrased positively ("ensure sharp focus on the subject's eyes" rather than "no blurry"), simulated identity consistency, and precise text integration using double quotes for legible signage. A prompt-optimization protocol requires the agent to rewrite the user's input into a logic-driven reasoning brief before invoking scripts/generate-nano-art.sh, which is called with flags such as --subject, --action, --context, and --style. Guardrails mandate full descriptive sentences over keyword soup, physically plausible lighting and reflections, and relationship-based phrasing. The skill functions as a logic wrapper around a core/media/generate-image.sh primitive.

Target users are creators, marketers, and agents who want cinematic, photorealistic, or precisely composed images with reliable text rendering. Typical use cases include product and scene imagery, stylized concept art, and images requiring accurate embedded text.