banana-cog
por nitishgargiitd
Coding Agents & IDEsResumen
Banana Cog × CellCog.
Descripción
Qué hace esta skill
Banana Cog wraps Nano Banana image generation with CellCog's orchestration layer to handle multi-image jobs from a single prompt. It maintains character consistency across sequences, plans scene compositions before generation, and can produce up to 30 coherent images in one request.
Calling Nano Banana directly produces one image at a time with no cross-image consistency; CellCog plans scenes upfront and holds character identity across the full batch.
Cuándo usarla
- Generating a consistent brand mascot across 10 different poses and contexts
- Building a storyboard sequence for a short animation pitch
- Producing a product photo series with matching lighting and style
- Illustrating a character through multiple story beats in one request
- Creating lifestyle mockups by placing products into new environments
Flujo de trabajo de ejemplo
User asks: Create a tech startup founder character and show them in 5 different workplace scenarios
- 1
CellCog's reasoning layer analyzes the request and plans 5 distinct scenes
- 2
Character design parameters are established to anchor consistent identity
- 3
Image generation calls are made to Nano Banana for each scene with shared character context
- 4
Consistency verification checks character identity across all outputs
- 5
Final set of 5 coherent images is delivered
5 images of the same character across varied workplace settings, visually consistent in appearance and style
Instalación
clawhub install nitishgargiitd/banana-cogEjecuta el comando de la CLI o pega el prompt en Claude Code, Cursor, Codex o el agente de programación que uses.