$ loading_
Generate revised images from annotated Cowart screenshots and place them nearby.
The materials indicate a screenshot-driven image-editing prompt workflow with no declared secrets, remote endpoints, or extra system permissions. Given that it is classified as prompt-only and open-source, the overall risk is low, though basic supply-chain verification is still advisable due to weak evidence of adoption and maintenance.
The materials explicitly state that no keys or environment variables are required, and the README does not request API tokens, account credentials, or sensitive local configuration. No clear credential leakage or abuse surface is evident.
The package declares no remote endpoint hosts, and the documentation does not describe uploading screenshots, images, or project data to third-party services. Based on the available materials, there is no explicit outbound data-transfer path.
The system flags it as prompt-only. The README mainly describes editing workflow and usage constraints, with no requirement to run shell commands, install extra binaries, or request local execution privileges beyond the stated image-editing purpose.
The materials emphasize deriving edit intent only from user-provided screenshots and explicitly say not to auto-capture the current canvas or scan the whole canvas. There is no indication of access to unrelated files, full-disk data, or excessive authorization.
The source is an open-source GitHub repository, which is a meaningful risk-reducing factor because the code is, in principle, auditable. However, the repository has 0 stars, no declared license, and unknown maintenance status, so evidence of community validation and ongoing upkeep is limited; basic repository review is recommended before adoption.
Copy the install command and let the AI configure it · recommended for beginners
Please install the "cowart-image-edit" skill from askskill: 1. Download https://raw.githubusercontent.com/zhongerxin/Cowart/main/skills/cowart-image-edit/SKILL.md 2. Save it as ~/.claude/skills/cowart-image-edit/SKILL.md 3. Reload skills and tell me it's ready
Designers or marketers can mark requested changes in Cowart with arrows and notes, then use the screenshot to generate a new bitmap revision. The original stays unchanged, and the revised result appears nearby for easy comparison.
When users provide multiple annotated screenshots, this skill processes each one as a separate task. It is useful for review rounds where several assets need edits without mixing requests across screenshots.
If a screenshot is too cropped, obstructed, or low resolution, the user can provide the original image or a cleaner screenshot. The skill uses the screenshot as the edit brief while preserving the original subject, composition, and style where possible.
The README explains how to turn user-supplied Cowart annotation screenshots into revised AI-generated bitmap images placed next to the originals. It covers prerequisites, the rule that screenshots are the authoritative edit brief, how to extract instructions from arrows and notes, when to ask for a cleaner source image, and how to generate a new output file without overwriting the original while carefully resolving the actual saved path.
Use this skill to turn user-provided Cowart 批注 screenshots into revised AI-generated bitmaps placed next to the corresponding original images.
The native Cowart widget should be open for the active project. Cowart state is read and written through Cowart MCP tools, not through a localhost browser service.
The user is responsible for providing the relevant screenshot(s). Do not auto-capture the current canvas and do not scan the whole canvas to infer edit requests; a canvas may contain many images with different annotations.
Read the user-provided screenshot(s).
Treat each screenshot as the authoritative edit brief for one output image unless the user says multiple screenshots belong to the same image.
If the user provides multiple screenshots, process them independently and keep their generated outputs separate. Do not merge annotations across screenshots unless explicitly requested.
Extract the edit requirements from each screenshot.
Read visible 批注 labels, arrows, and nearby edit notes from the screenshot itself. Use the arrow tip or marked region to understand where each note applies.
Ignore editor chrome such as toolbars, blue selection outlines, resize handles, cursor icons, and unrelated neighboring images.
Choose the source image for generation.
Use the clean underlying image content visible in the provided screenshot as the visual base whenever possible.
If the screenshot is too cropped, obstructed, or low-resolution to serve as a good image base, ask the user for the original image export or a cleaner screenshot of that specific image.
Do not read the current Cowart canvas to discover edit intent. Use the screenshot for the requested changes. Cowart state may be read later only to place the generated result without covering existing content.
Prepare image-generation input.
Use the provided screenshot, plus a cleaner source image if the user supplied one.
The generation prompt should:
Generate a new bitmap.
Use the built-in image generation flow available in the current environment. Do not overwrite the source image file. Save the new bitmap with a timestamped filename, for example:
annotation-edit-20260620-153012.png
Resolve the actual local output image carefully before inserting it into Cowart. Do not assume the built-in image generation flow always writes a fresh file under $CODEX_HOME/generated_images.
Preferred resolution order:
image_generation_call.result, then write it to the timestamped output filename.$CODEX_HOME/generated_images only when you can prove the file was created by the current request, for example by matching its timestamp after this generation step. Never pick an older image merely because it is the newest file in a stale generated_images directory.Before inserting the resolved file into Cowart, visually inspect the local bitmap and confirm it is the newly generated revised image for this screenshot, not a stale generated asset.
Insert the revised image beside the original with Cowart MCP.
Prefer the Cowart MCP insert_cowart_image tool. Do not hand-write
tldraw asset / shape records or fractional index keys unless the MCP
tool is unavailable. The tool copies the bitmap into the page-local assets
…
It reads user-provided Cowart annotation screenshots, treats the visible arrows, labels, and notes as edit instructions, and generates a clean revised bitmap. The result is placed beside the corresponding original or in a nearby clear area without replacing or hiding the original.
The documentation says the native Cowart widget for the active project should already be open, and Cowart state is read and written through Cowart MCP tools. Users must provide the relevant screenshots; the skill does not auto-capture the canvas or scan the whole canvas to infer edits.
No, not by default. The documentation says each screenshot should be treated as the authoritative edit brief for one output image unless the user explicitly says multiple screenshots belong to the same image.
Generate or edit bitmap images for illustrations, photos, textures, mockups, and cutouts.
Analyze design images for layout, color, typography, and accessibility improvements.
Create original posters, illustrations, and static visuals in PNG or PDF.
Generate or edit images from natural language with CCGO image MCP.
Enhance image quality, sharpness, and resolution for docs, presentations, and social posts.
Generate images from text prompts for fast visual ideation in conversations.