Image Proxy

Generate or edit raster images through an explicitly configured OpenAI-compatible third-party image provider, including ThinkAI by default, with dry-run request snapshots, source-image and mask uploads, progress files, secure local key configuration, and downloaded output verification. Use when the user asks to use a third-party image proxy, ThinkAI, an OpenAI-compatible image endpoint, or the image-proxy skill; do not use it for native Codex image generation or a full ecommerce production workflow.

43 stars 5 forksImages, Video & Audio

Install

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add ninemouth/sellerpilot-product-image-industrial --skill image-proxy
Manual install
git clone --depth 1 https://github.com/ninemouth/sellerpilot-product-image-industrial.git /tmp/sellerpilot-product-image-industrial
mkdir -p ~/.claude/skills
cp -r /tmp/sellerpilot-product-image-industrial/standalone/image-proxy ~/.claude/skills/image-proxy

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# Image Proxy

Use this skill as an independent provider execution layer. It sends a finalized prompt to one configured OpenAI-compatible image endpoint, optionally uploads source/reference images and a PNG mask for edits, downloads the returned image assets, and writes request/response/summary evidence in the requested output directory.

## Codex chat commands

Users operate this Skill from Codex chat. Do not require them to open Terminal or paste an API Key into the conversation.

```text
$image-proxy status
$image-proxy dry-run --prompt "<final prompt>" --size 2k --quality hd
$image-proxy generate --prompt "<final prompt>" --size 2k --quality hd
$image-proxy edit --prompt "<final prompt>" --image <attached-or-local-image> --mask <optional-png-mask>
$image-proxy sync
```

- `status` resolves only non-secret local provider state.
- `dry-run`, `generate`, and `edit` are Codex intents. Codex runs the deterministic scripts internally and returns the output/evidence; the user should not be sent to a shell.
- A Web-managed installation automatically pulls and applies the current user's `sellerpilot-image` configuration before every real third-party request. If the shared device session is missing, the tool opens the Marqel Web approval flow, waits for the current user to approve it, and resumes the same configuration pull automatically. Dry-runs stay network-free and never trigger configuration delivery.
Read the full skill on GitHub

Details

Skill name
image-proxy
Path
standalone/image-proxy/SKILL.md
License
MIT
Repo updated
3 September 2026

FAQ

What does the Image Proxy skill do?

Generate or edit raster images through an explicitly configured OpenAI-compatible third-party image provider, including ThinkAI by default, with dry-run request snapshots, source-image and mask uploads, progress files, secure local key configuration, and downloaded output verification. Use when the user asks to use a third-party image proxy, ThinkAI, an OpenAI-compatible image endpoint, or the image-proxy skill; do not use it for native Codex image generation or a full ecommerce production workflow.

How do I install the Image Proxy skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add ninemouth/sellerpilot-product-image-industrial --skill image-proxy. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Image Proxy popular?

It lives in ninemouth/sellerpilot-product-image-industrial, which has 43 GitHub stars and 5 forks.

Related skills

Sellerpilot Product Image Industrial

43

ninemouth/sellerpilot-product-image-industrial

Create, plan, review, revise, and audit industrial ecommerce product-image sets with source-identity preservation, platform-aware composition, localized copy, provider-safe generation, QA routing, tldraw review, and run-scoped delivery.

Images, Video & Audio

Media Scripter

50

YujxZJCN/teaching-skills

Recorded-media scripting for university professors. 4-agent team turning lecture notes and flipped-class specs into mini-lecture video scripts, shot-by-shot storyboards, 6–9-minute episode series, cleaned captions/transcripts, and podcast-style audio adaptations. Scripts are spoken language with built-in accessibility — not read-aloud prose. Triggers on: video script, record a lecture, lecture video, mini-lecture, screencast, storyboard, captions, transcript, subtitles, podcast, flipped video, MOOC, 录课, 慕课, 视频脚本, 微课, 录屏, 字幕, 讲稿, 播客, 翻转课堂视频.

Images, Video & AudioIn 1 list

Scoop

58

EricTechPro/startup-claude-skills

Competitor content gap analysis — extracts ALL comments from competitor and your own YouTube videos, identifies what audiences want but aren't getting, ranks gaps by engagement, and proposes video ideas. Use when the user says "scoop," "content gap," "analyze competitor videos," "what should my next video be about," "find content gaps," or "competitor analysis.

Images, Video & Audio18

Plotter

46

pasqualevitiello/plotter

Draw, critique, refine, or animate flat monochrome blueprint illustrations as editable SVG — thin technical line drawings that explain an idea through one construction relationship. Use when the user wants an illustration, hero image, spot illustration, empty-state or 404 image, or an explanatory drawing for docs, a landing page, a feature page, or a README in a blueprint, technical-drawing, schematic, or line-art style; when they name Plotter or a catalogue code such as TIM-05; or when they ask to animate or review a Plotter drawing. Not for data charts, flowcharts in Mermaid or similar tools

Images, Video & Audio

Olore Recraft Latest

119

olorehq/olore

Local recraft documentation reference (latest). Recraft image generation API and Studio. Use for AI image generation, vector/SVG logos, prompt engineering, styles, image editing, vectorization, and background removal.

Images, Video & Audio

Multi Style Image Generator

44

shengjidaguai-china/multi-style-image-generator

Generate or stylize images and prompts in game, fantasy, pixel or crochet/yarn styles; derive spatial depth, single-image point clouds, 360° panoramas, explorable panorama point-cloud scenes and CogVideoX videos. Use for 出图、照片风格化、毛线编织、空间景深图、点云、360° 全景图、360° 点云漫游 or video requests.

Images, Video & Audio

Browse the most popular Claude skills, all images, video & audio skills, or write your own with the skills generator.