nano-banana-edit
Edit Images with Google Nano Banana 2 on RunComfy
Auch verfügbar von: runcomfy-com,doany-ai
AI image editing typically requires complex API calls and batch management. This skill provides Claude with complete knowledge of the Google Nano Banana 2 image-to-image edit endpoint, including prompt engineering best practices, batch processing for up to 20 images, and routing guidance for when to use alternative models.
Die Skill-ZIP herunterladen
In Claude hochladen
Gehe zu Einstellungen → Fähigkeiten → Skills → Skill hochladen
Einschalten und loslegen
Teste es
Verwendung von "nano-banana-edit". User: Edit this product photo with a minimalist white background
Erwartetes Ergebnis:
The skill invokes the RunComfy CLI to call google/nano-banana-2/edit with a prompt structured as: 'Keep the product identity, shape, and labeling unchanged. Convert the background to a clean white minimalist studio backdrop.' The model returns the edited image saved to the specified output directory.
Verwendung von "nano-banana-edit". User: Remove the person on the left from all three photos, keep the right side exactly the same
Erwartetes Ergebnis:
The skill calls the edit endpoint with three image URLs and a spatial prompt: 'Remove the leftmost person only. Keep the right side of the image, the background, and all other elements exactly as in the input.' Three edited images are returned with consistent edits applied.
Sicherheitsaudit
Niedriges RisikoAll static findings are false positives. The skill is a documentation wrapper that calls the RunComfy CLI to invoke the Google Nano Banana 2 image-to-image edit endpoint. Commands shown in documentation (SKILL.md) are markdown code blocks, not actual shell execution. The CLI passes user prompts as JSON via the --input flag, not via shell string interpolation, eliminating command injection risk. The skill properly documents its security architecture including token storage, input boundary handling, and network endpoints.
Risikofaktoren
⚙️ Externe Befehle (5)
🌐 Netzwerkzugriff (4)
📁 Dateisystemzugriff (4)
Qualitätsbewertung
Was du bauen kannst
E-commerce product catalog updates
Swap backgrounds on product images while preserving the product identity, consistent across an entire SKU catalog.
Marketing creative variations
Generate A/B variants of ad creative by editing backgrounds or adding localized elements while maintaining brand consistency.
Influencer content localization
Replace backgrounds on influencer spokesperson photos for different markets while preserving identity across edits.
Probiere diese Prompts
Keep the subject identity unchanged. Convert the background into a rainy neon cyberpunk street.
Remove the leftmost object only. Keep the right two objects, the table, and the lighting unchanged.
Keep the bottle, label, and lighting exactly as in the input. Replace only the brand text on the label from "ALPHA" to "AURA", same font weight, centered, white on black.
For each input image: keep the subject's pose and identity unchanged. Convert the background to a soft warm-grey studio sweep with subtle floor shadow. Center the subject at the same fraction of frame as the input.
Bewährte Verfahren
- Lead prompts with preservation goals before stating changes; models honor what is stated up front
- Use concrete spatial language (upper-left, bottom-right) rather than vague terms like 'more modern'
- Lock aspect_ratio and resolution when batch editing to ensure consistent framing across outputs
Vermeiden
- Compound instructions with multiple unrelated changes cause drift; split into sequential passes
- Passive voice edits like 'the background should be changed' reduce effectiveness; use imperative mood
- Missing preservation goals allows the model to subtly modify faces or brand elements unintentionally