GitHub agent generates App Store screenshots
A GitHub project demoed AI agents that generate App Store screenshots at Apple resolutions, offering a new on‑device ML workflow to help iOS developers produce store assets quickly shared. The tool surfaced in social channels as a practical example of agentic automation for developer UX tasks.
ParthJadhav’s GitHub skill scaffolds a Next.js generator and explicitly exports PNGs at four Apple target sizes, calling out 6.9", 6.5", 6.3" and 6.1" outputs. (github.com) A fork by keremerkan adds device framing and a zip-export that’s declared compatible with asc-client for direct App Store Connect uploads. (github.com) The ParthJadhav README lists concrete pixel targets—1320×2868 for the largest 6.9" canvas, then 1284×2778, 1206×2622 and 1125×2436 for smaller devices. (github.com) Multiple public skill indexes and READMEs advertise that these generators are packaged as “skills” for agent platforms such as Claude Code, Cursor, Windsurf, OpenCode and Codex. (skillstore.io) Several forks and READMEs include operational notes for CI and capture workflows, for example a README that instructs using the 6.1‑inch simulator as the canonical source capture to avoid downstream resizing adjustments. (github.com) keremerkan’s implementation documents a macOS dependency on FrameMe for pixel‑perfect bezels (with swift build steps) and positions its output to auto-scale from the largest Apple-required size down to device variants. (github.com)
