Stats
Actions
Tags
From ImageKit Plugin
MANDATORY PRE-STEP: You MUST read and follow this skill BEFORE calling the `transformation_builder` tool. This skill teaches you how to identify the correct ImageKit capability, craft precise queries, and handle multi-step transformations. Covers AI editing (change objects, colors, styles), background removal/replacement, generative fill, upscaling, retouching, resize, crop, overlays, text overlays, blur, sharpen, rotate, borders, shadows, color replace, and all visual modifications. Never call the transformation_builder MCP tool without first consulting this skill.
How this skill is triggered — by the user, by Claude, or both
Slash command
/imagekit-plugin:transformation-builderThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
This skill MUST be invoked and read before you call `transformation_builder`. It contains critical instructions for identifying the right capability, crafting precise queries, and ordering multi-step transformations that determine whether the tool produces correct results.
This skill MUST be invoked and read before you call transformation_builder. It contains critical instructions for identifying the right capability, crafting precise queries, and ordering multi-step transformations that determine whether the tool produces correct results.
Workflow:
transformation_builder with the prepared queryWhen user says "remove background" without specifying the approach, ask which option they prefer:
e-bgremove via transformation URL. Best for on-the-fly delivery, no new file stored.Only apply ai_changebg if the user wants to replace the background with a new scene or design. If they just want to remove the background and change color or keep it transparent, use or ai_remove_background or ai_removedotbg instead.
ALWAYS apply background removal AFTER upscale/retouch, NEVER before.
Correct: Upscale -> Then remove background
Wrong: Remove background -> Then upscale
Use e-bgremove (not ai_remove_background) for background removal.
transformation_builderAfter reading this skill and preparing your query, call the tool with:
query: The precise, rewritten query (see Query Crafting Guide below — NEVER pass vague user input directly)src: Optional source URL (defaults to sample image)Chain transformations: Frame multi-step requests as a single query.
Users describe what they want in vague, everyday language. Your job is to identify the right IK capability and rewrite the query into a precise, tool-friendly description. Never pass vague queries through as-is.
Use this table to decide which capability handles the user's request:
| User Intent (vague) | Maps To | Why |
|---|---|---|
| "Make the red balls green", "enlarge the cat", "remove the car", "add sunglasses" | ai_edit | Modifying specific objects or regions within the image requires AI content editing |
| "Put this on a beach", "change background to sunset", "make it look like they're in Paris" | ai_changebg | Replacing the background scene while keeping the subject. Adds new scene elements defined by prompt. Should not be used for simple background removal or color changes. |
| "Remove the background", "make it transparent", "cut out the person" | ai_remove_background | Removing the background entirely (transparent output) |
| "Make it wider without stretching", "extend the sky", "add more space on the left" | ai_bg_genfill | Expanding the image canvas with AI-generated content beyond original bounds |
| "Make it higher resolution", "it's blurry can you fix it", "enhance quality" | ai_upscale | Increasing resolution / fixing low-res images |
| "Clean up this photo", "remove blemishes", "fix skin imperfections" | ai_retouch | General visual quality improvement and artifact removal |
| "Add a realistic shadow", "make the product look like it's on a surface" | ai_drop_shadow | Adding natural-looking shadows (transparent images only) |
| "Create a variation", "give me a different version", "remix this image" | ai_genvar | Generating visual variations while preserving structure |
| "Generate an image of a sunset over mountains" | ai_gen_image | Creating a brand new image from text (no source image needed) |
| "Make it 800x600", "resize for Instagram", "make it smaller", "crop to 16:9" | resize_and_crop | Changing overall dimensions, aspect ratio, or crop of the image |
| "Focus on the face", "crop around the product", "zoom into the subject" | resize_and_crop (with focus) | Smart cropping using face/object detection |
| "Change all red pixels to blue", "swap the background color from white to gray" | color_replace (effects) | Global pixel-level color swap across the entire image |
| "Make it black and white", "blur it", "sharpen", "rotate 90°", "add a border" | effects_and_enhancement | Standard image filters and adjustments |
| "Add my logo on top", "put a watermark", "overlay this badge" | image_overlay | Compositing another image on top |
| "Write 'SALE 50% OFF' on the image", "add a caption" | text_overlay | Rendering text on the image |
ai_edit vs color_replace vs ai_changebgThese three are commonly confused. Use this decision tree:
ai_editcolor_replaceai_changebgRule: Always rewrite the user's vague request into a query that references actual IK parameters and capabilities. Use parameter names when you know them; fall back to descriptive English only when the exact parameter is unclear.
| User Says | Bad Query (don't do this) | Good Query |
|---|---|---|
| "Make the red balls bigger and the blue ones green" | "make red balls bigger and blue balls green" | "Apply ai_edit with prompt: increase the size of the red-colored balls and change the blue-colored balls to green" |
| "Can you clean this up and put it on a white background?" | "clean up and white background" | "Apply ai_retouch, then ai_remove_background, then ai_changebg with prompt: solid white background" |
| "I want this to look like a painting" | "make it look like a painting" | "Apply ai_edit with prompt: transform into an oil painting style with visible brush strokes and artistic texture" |
| "Crop around the person's face" | "crop face" | "Apply resize_and_crop with focus set to face to crop around the detected face" |
| "Add some text saying Hello World at the bottom" | "add hello world text" | "Apply text_overlay with text='Hello World', layer_y=bottom, layer_focus=bottom" |
| "Make this fit for a mobile banner" | "mobile banner" | "Apply resize_and_crop with width=640, aspect_ratio=2-1" |
| "The background is boring, make it interesting" | "change background" | "Apply ai_changebg with prompt: vibrant gradient studio backdrop" |
| "Extend the image to make it panoramic" | "make panoramic" | "Apply ai_bg_genfill to extend image horizontally with aspect_ratio=3-1" |
| "This is too low quality" | "fix quality" | "Apply ai_upscale, then ai_retouch" |
| "Make it 300px wide and add a red border" | "resize and border" | "Apply resize_and_crop with width=300, then apply border with color red and width 5" |
| "Blur everything and put white text on top" | "blur and text" | "Apply blur=10, then text_overlay with text='Your Text', color=white, font_size=40, layer_focus=center" |
| "Remove background and add a drop shadow" | "remove bg shadow" | "Apply ai_remove_background, then ai_drop_shadow" |
When the user's request involves multiple capabilities, chain them in one query in the correct order:
Example: User says "clean up this photo, remove the background, and make it 500x500"
→ Query: "Apply ai_retouch, then ai_remove_background, then resize_and_crop with width=500, height=500"
search-docs skill first, then call search_docs for parameters/limitssearch-docs skill, then call search_docs to find supported methods and constraintsresize_and_crop)| IK Parameter | What It Does |
|---|---|
width | Output width. If only width is provided, height auto-scales to preserve aspect ratio. Accepts integer (px), decimal 0–1 (percentage), or arithmetic expression. |
height | Output height. If only height is provided, width auto-scales to preserve aspect ratio. Accepts integer (px), decimal 0–1 (percentage), or arithmetic expression. |
aspect_ratio | Sets aspect ratio (width:height). Must be used with either width or height; ignored if both are provided. |
crop_mode | Controls resize/crop behavior: pad_resize (fit within dimensions with padding), extract (extract fixed region), pad_extract (padded extract). |
crop | Controls crop strategy: force (exact resize, ignores aspect ratio), at_max_enlarge (resize at max without enlarging), at_least (cover minimum dimensions), maintain_ratio (preserve ratio). |
focus | Focal point for cropping. Values: custom (predefined area), auto (smart crop), face (face detection), or an object name for object detection. |
zoom | Controls zoom level around the focused area during face/object-based cropping. |
x, y, x_center, y_center | Coordinates for cm_extract: x/y set top-left corner; x_center/y_center set center of extracted region. |
dpr | Device pixel ratio — scales output dimensions for high-DPR displays (e.g. Retina). |
background | Background color for padded areas. Supports hex, named colors, blurred background, and dominant color modes. |
ai_transform)| IK Parameter | What It Does |
|---|---|
ai_remove_background_external | Removes image background using external AI provider (remove.bg), producing transparent background. |
ai_remove_background | Removes image background using ImageKit native AI, producing transparent background. |
ai_drop_shadow | Adds AI-generated drop shadow around main subject. Controls: light direction, elevation, shadow intensity. Works on transparent images only. |
ai_changebg | Changes image background using a text prompt while preserving the foreground subject. Supports plain text and base64-encoded prompts. |
ai_edit | Edits/modifies image content using a descriptive text prompt with AI-driven visual changes. |
ai_bg_genfill | Extends image beyond original boundaries by generating new background content using AI (generative fill / outpainting). |
ai_gen_image | Generates a new image from a text prompt using AI and stores it at a specified file path. |
ai_genvar | Generates AI-based visual variations of an existing image while preserving overall structure and composition. |
ai_retouch | Improves visual quality using AI — enhances details and corrects imperfections. No additional parameters. |
ai_upscale | Increases image resolution using AI super-resolution upscaling. No additional parameters. |
image_overlay)| IK Parameter | What It Does |
|---|---|
image_path | Path of the overlay image from ImageKit media library. Can also specify a solid color for color block overlays. |
encoded | Whether to base64-encode the image path. |
layer_x | Horizontal position of the overlay relative to the base image. Supports arithmetic expressions and negative offsets. |
layer_y | Vertical position of the overlay relative to the base image. Supports arithmetic expressions and negative offsets. |
height | Height of the overlay image. |
width | Width of the overlay image. |
crop_mode | Controls padding vs extraction behavior for the overlay. |
background | Background color for padded areas of the overlay. |
layer_focus | Positions the overlay relative to its parent container (e.g. center, top_left, bottom_right). |
layer_mode | Blending mode: multiply (darken blend), cutout (cut transparent regions), cutter (mask base using overlay shape). |
effects.grayscale | Converts overlay to grayscale. |
effects.border | Adds border around the overlay. |
effects.trim | Auto-trims transparent edges of overlay. |
effects.zoom | Zooms into cropped region of overlay. |
x, y, xc, yc | Defines exact crop region using absolute or center-based coordinates. |
text_overlay)| IK Parameter | What It Does |
|---|---|
text | The text string to overlay on the image. |
layer_x, layer_y | Horizontal and vertical position of the text layer relative to the base image. |
width | Maximum width of the text box. Text wraps automatically when exceeded. |
font_size | Font size of the overlaid text. Supports numbers and arithmetic expressions. |
color | Text color. Supports named colors, RGB hex, and RGBA hex with opacity. |
inner_alignment | Horizontal alignment of text within the text box. |
padding | Space around the text inside its background box. |
alpha | Transparency of the entire text layer. |
typography | Typographic styling: bold, italic, strikethrough, or combinations. |
font_family | Font used to render the text. |
line_height | Spacing between lines when text wraps across multiple lines. |
flip | Flips or mirrors the text layer horizontally/vertically. |
layer_mode | Blending mode for text: multiply, cutout, or cutter. |
effects_and_enhancement)| IK Parameter | What It Does |
|---|---|
contrast | Auto-enhances contrast by stretching pixel intensity to full range. |
sharpen | Basic sharpening. Excessive values may cause artifacts. |
unsharp_mask | Advanced sharpening via Unsharp Masking (USM) with better perceptual quality. |
shadow | Adds drop shadow beneath non-transparent pixels (requires transparent background). |
gradient | Applies linear gradient overlay/background with configurable colors, direction, and stop point. |
grayscale | Converts image to grayscale. |
perspective_distort | Distorts image perspective using custom coordinates. |
arc_distort | Curves the image upward or downward. |
trim | Removes solid/near-solid background pixels around a central object. |
blur | Applies Gaussian blur with configurable intensity. |
border | Adds border with specified width and color around the image. |
rotate | Rotates image clockwise/counter-clockwise or auto-rotates using EXIF data. |
flip | Mirrors image horizontally, vertically, or both. |
radius | Rounds image corners; max value produces a fully circular image. |
color_replace | Replaces a source color and similar shades with a target color, preserving luminosity. |
video_transforms)| IK Parameter | What It Does |
|---|---|
radius | Rounds video corners. Use max for circular/oval clipping. |
rotate | Rotates video clockwise by a fixed degree. |
mute_audio | Mutes the video audio track. |
extract_audio | Extracts audio track and outputs as audio-only file. |
border | Adds border around the video frame with configurable width and color. |
trim_video | Trims video using start_time, end_time, and duration parameters. Supports arithmetic expressions. |
video_overlay)| IK Parameter | What It Does |
|---|---|
image_path | Path of image to overlay on video. Supports base64-encoded paths via encoded. |
text | Text string to overlay on video. |
video_path | Path of video to overlay on base video. |
background | Solid color for color block overlays on video. Supports named colors, RGB hex, RGBA hex. |
subtitle_path | Path to subtitle file for subtitle overlays. |
layer_x, layer_y | Position of any overlay (image/text/video/solid) relative to the base video. |
layer_focus | Aligns overlay relative to base video frame (e.g. center, top_left). |
start, end, duration | Controls overlay timing — when it appears, disappears, and how long it stays visible. |
color | Text color for text overlays. Supports named colors, RGB hex, RGBA hex. |
font_size | Font size for text overlays. Supports numbers and arithmetic expressions. |
typography | Typography style for text overlays. |
npx claudepluginhub imagekit-developer/skills --plugin imagekit-pluginFetches up-to-date documentation from Context7 for libraries and frameworks like React, Next.js, Prisma. Use for setup questions, API references, and code examples.
Applies a firm's KYC/AML rules grid to parsed onboarding records: assigns risk rating, checks required documents, outputs rule outcomes with citations, and routes for escalation.
Generates daily or weekly digests of activity from connected sources (chat, email, docs, tasks, CRM), highlighting action items, decisions, mentions, and project updates.