web-infra-dev
2 skills in one repo.
midscene-skills Skill collection
Vision-driven cross-platform automation with Midscene.js: skills for browsers, desktops, Android, iOS and HarmonyOS.
320 stars on GitHub, Nov 9, 2026 · 2 skills
-
- Use it when:
- You want the agent to operate or test an app on an Android phone or emulator, without its code or accessibility labels.
- It does:
- Runs one Midscene command at a time: connect, launch an app, act on a plain-language goal, take a screenshot, assert what the screen shows.
- Age
- 8 months
- Stars
- 320
- bottom 9%
- Total Downloads
- 5.5K
- bottom 16%
- Weekly Downloads
- 70
- bottom 13%
-
- Use it when:
- You want the agent to browse, fill forms, upload files, scrape data or test a web page by what it looks like, not its DOM.
- It does:
- Runs one Midscene command at a time: connect to a page, act on a plain-language goal, take a screenshot, assert what the page shows.
- Age
- 8 months
- Stars
- 320
- bottom 9%
- Total Downloads
- 5.4K
- bottom 13%
- Weekly Downloads
- 58
- bottom 12%