edit-image
Skill4 starsChanged 4 months agoMITView on GitHub
Use when the user wants to remove, blur, or replace an image background, remove an object, fill a masked region, extend a canvas, restore a photo, upscale, or convert an existing image in Simplified. For a new image or broad reference-guided transformation, use generate-image.
3 other projects use a version of this.
Adopted by
Most recently taken up by
- Runware/runware-skillstheir own version, first seen Oct 15, 2026
- Sateezg/codex-bridgetheir own version, first seen Aug 9, 2026
- SamurAIGPT/muapi-claude-codetheir own version, first seen Jun 11, 2026
Other versions3 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 3
- Sateezg/codex-bridge★ 248same name
- Runware/runware-skills★ 8same name
- SamurAIGPT/muapi-claude-code★ 4same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository23 other files it gives its agents.
- AGENTS.mdAGENTS.md
- campaign-reviewSkill
- content-repurposerSkill
- creative-testingSkill
- cross-platform-campaignSkill
- evergreen-content-engineSkill
- generate-imageSkill
- generate-videoSkill
- generate-voiceoverSkill
- local-business-marketingSkill
- manage-assetsSkill
- manage-brandSkill
- manage-projectsSkill
- repurpose-videoSkill
- simplified-agentsSkill
- simplified-cliSkill
- simplified-project-managementSkill
- simplified-socialSkill
- simplified-workflowsSkill
- simplified-workspaceSkill
- social-content-plannerSkill
- social-performance-analystSkill
- transcribe-mediaSkill
The file
We split this skill into the rules it gives an agent
1 rule. 0 are common, 1 is distinctive, and 0 are very specific to this project.
0
Common
Most of the 3 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 12
Use the connected Simplified tools and their current schemas.
“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Edit an existing image
Choose the smallest supported operation that achieves the requested edit. Preserve the source and create a new result; do not regenerate the product or change approved copy to solve a file-processing task.
Resolve the source and scope
Use the connected Simplified tools and their current schemas. Canonical names below may have a client prefix. A missing tool is a capability limit, not a reason to silently switch providers.
For a named or uncertain workspace/teamspace, resolve it with api_getWorkspaceInfo and accessible membership from api_listTeamspaces. Carry the numeric space_id on every related call. Use manage-assets to search names/tags or handle intake; a folder name does not establish its contents.
For an asset UUID, call api_getAsset, require ready status: 4, and use its fresh full file URL. The image editing tools below take image_url, not an asset UUID or a client-local path. An attachment must first be made accessible using a supported import/upload flow. Preserve signed URL query strings and refresh expired source URLs.
Choose the operation
| Requested change | Operation and inputs |
| Transparent cutout | media_removeBackground: image_url, output_format: "png"; omit background_color; avoid magic_crop: true when framing must remain |
| Background blur | media_blurBackground: image_url, required blur_value from 1–100 |
| Background color/image/transparency | media_replaceImageBackground: replace_type; color needs replace_color, image needs replace_image URL |
| Resolution | media_upscaleImage: scale 2, 4, or 8, according to request |
| Denoising/artifact repair | media_restoreImage; use its supported strength fields |
| Specific masked region | media_generativeFill: prompt and mask_url or mask_base64; white marks the fill area |
| Object described in words, no mask | media_magicInpaint: prompt describing removal/replacement |
| Extend beyond existing borders | media_imageOutpainting: prompt and required mask_url defining the extension |
| File format only | media_convertImageFormat: current format field and supported value |
For an exact-region request without a mask, obtain or create a usable mask only if the client can do so accurately. Semantic inpainting is an alternative only when the user accepts model-selected boundaries. Neither a prompt nor a mask guarantees pixel-perfect preservation: verify the result, and do not promise unchanged pixels outside a region without comparison.
Use PNG for transparency; JPEG cannot retain alpha. Do not promise that later upscaling or conversion retains transparency or source framing without inspecting it. If visual inspection is unavailable, report that limit. Masked generation can default to multiple variations: set count: 1 unless more were requested. Explicit edits authorize the requested work; clarify materially ambiguous changes or added variants before spending credits.
Complete each step before chaining
media_blurBackground can return an image_url synchronously. Other tools may be async at the backend while middleware waits for completion. Inspect the actual completed output; a task_id alone is not an image.
For a pending/timeout envelope, retain the exact task ID and call api_getTaskResult where exposed. Poll with reasonable intervals and a bounded wait. SUCCESS requires usable output; FAILURE or REVOKED stops the chain. A timeout with error: true is incomplete, not proof of terminal failure. Do not resubmit the paid edit or run the next step on the original as if the previous edit succeeded. If no continuation tool is exposed, report the saved task ID and unfinished steps.
For remove-background → upscale, the upscale input is the completed cutout URL. For a batch, preserve successes and report each failure; retry only unresolved work when its state and authorization are clear.
Save and hand off
If the edit returns a verified permanent asset UUID, reuse it rather than importing a duplicate. If it returns only a usable temporary URL and the result should be retained, call api_createAsset with that URL and a useful name, then api_getAsset until ready. Never regenerate just to obtain storage. The original asset remains separate.
Return the edit performed, actual result link, verified asset UUID/readiness, and any unverified transparency/framing detail. Show links in the form supported by the client. For requested social drafts, pass the ready result UUID to simplified-social with action: "draft"; editing does not authorize publication.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

