agentleFS
yanzyuyuyanzyuyu/min-tok/skills/omni-pilot/SKILL.md

omni-pilot

Skill1 starsChanged 49 days agoApache-2.0View on GitHub

MANDATORY: Execute whenever the user requests full OS control, GUI automation outside of browsers, or bypassing HTML DOM scripting. Uses PyAutoGUI and ImageGrab to physically move the mouse and type based on raw visual coordinates.

The file

Files
SKILL.md

Omni-Pilot (Pixel-Level OS Embodiment)

Tujuan

Menghentikan ketergantungan AI pada kode HTML/DOM yang rapuh. Omni-Pilot memungkinkan AI "melihat" layar penuh (seluruh OS, bukan cuma Chrome) menggunakan Multimodal Vision, lalu menggerakkan mouse dan keyboard secara fisik layaknya manusia.

Workflow Eksekusi (Computer Use)

  1. Visual Capture: Jalankan perintah python C:\Users\MSI\.gemini\config\utils\omni_pilot.py --action capture. Ini akan menyimpan gambar layar di C:\Users\MSI\.gemini\antigravity\scratch\omni_screen.png.
  2. Visual Processing: Baca file omni_screen.png tersebut menggunakan alat view_file (Mata Multimodal Anda). Cari letak elemen/tombol yang ingin Anda klik (perkirakan koordinat X dan Y secara piksel dari kiri atas).
  3. Physical Click: Jalankan perintah python C:\Users\MSI\.gemini\config\utils\omni_pilot.py --action click --x [X_COORD] --y [Y_COORD] untuk menggerakkan kursor secara nyata dan mengekliknya.
  4. Physical Typing: Jika ingin mengetik, klik dulu area teksnya, lalu jalankan python C:\Users\MSI\.gemini\config\utils\omni_pilot.py --action type --text "teks anda".
  5. Physical Press: Jika ingin menekan Enter/Escape, jalankan python C:\Users\MSI\.gemini\config\utils\omni_pilot.py --action press --text "enter".

Aturan Absolut

  • BANNED (Puppeteer/DevTools): Anda DILARANG KERAS menggunakan puppeteer_evaluate, chrome-devtools, atau mencari DOM dengan skrip jika User mengaktifkan mode Omni-Pilot ini.
  • Fail-Safe: Jika kursor bergerak liar, arahkan segera koordinat X=0, Y=0 untuk memicu Abort.
  • Selalu capture ulang layar setiap kali tata letak GUI berubah (misal: setelah klik). Jangan menebak koordinat buta.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.