kortix computer

Skill

How to reach a CONNECTED MACHINE (a user's laptop/desktop, or any computer paired over the Agent Computer Tunnel) from a Kortix session — read/write files, run shell commands, and drive the desktop (click/type/screenshot) on that machine. It works through the Executor's `computer` connector (the same connectors/discover/describe/call path as every other integration), so there is no separate tunnel client and no token. Load this when the task is about acting ON a specific physical/remote computer the user has connected ("on my laptop…", "read ~/Downloads on my machine", "run this on my desktop", "click the button on my screen"), or when the user asks how the agent reaches their computer. For files INSIDE this sandbox, just use normal shell/fs — not this.

Files1
  • @skills/kortix-computer/SKILL.md
The computer connector's tools relay an RPC to the machine:
  • filesystem — computer.fs.read / fs.write / fs.list / fs.stat / fs.delete
  • shell — computer.shell.exec (stdout / stderr / exitCode)
  • desktop — computer.desktop.cua.click / type_text / press_key / hotkey / scroll / launch_app / list_apps / list_windows / get_screen_size / get_accessibility_tree, plus computer.desktop.cua.call (a passthrough to ANY computer-use tool by name).
Every relayed tool takes a computer argument selecting which machine (its name or id). It's optional when exactly one machine is online — then that one is used by default.
This is for a connected, external computer — not this sandbox. To touch files in your own workspace, use normal shell/fs. Reach for computer.* only when the task is explicitly about the user's own machine.
jsonc
{ "connector": "computer", "action": "list_computers" }
// → { "computers": [ { "id": "…", "name": "Marko's MacBook", "online": true,
//                      "capabilities": ["filesystem","shell","desktop"], "platform": "darwin" } ] }
If connectors doesn't list a computer connector at all, the user hasn't connected a machine — tell them to connect one in Customize → Computers (or kortix tunnel). If it's listed but list_computers shows the target online: false, ask them to bring it online.
2. Call a tool, picking the machine. Pass computer (name or id). Omit it when only one machine is online.
jsonc
// read a file on the laptop
{ "connector": "computer", "action": "fs.read",
  "args": { "computer": "Marko's MacBook", "path": "/Users/marko/notes.md" } }

// run a command (sole online machine → no selector needed)
{ "connector": "computer", "action": "shell.exec",
  "args": { "command": "git", "args": ["status"], "cwd": "/Users/marko/proj" } }

// drive the desktop
{ "connector": "computer", "action": "desktop.cua.type_text",
  "args": { "computer": "Marko's MacBook", "text": "hello" } }
describe any tool first if unsure of its inputs (e.g. { "tool": "computer.shell.exec" }).
3. The passthrough for the long tail. Beyond the curated desktop actions, computer.desktop.cua.call invokes any computer-use tool by name:
jsonc
{ "connector": "computer", "action": "desktop.cua.call",
  "args": { "computer": "Marko's MacBook", "tool": "double_click", "args": { "x": 220, "y": 140 } } }
If you call something that isn't yet granted, the call comes back as pending_approval: a permission request is created and surfaced to the user in Computers. Tell them what you're trying to do and that they need to approve the request in Computers, then retry once they have. Don't try to route around a denial (there is no token to fall back to, by design).