agentleFS
Sign inStart free
calebnewtonusccalebnewtonusc/Nova/skills/mac-act/SKILL.md

mac-act

Skill0 starsChanged 2 months agoMITView on GitHub

Click, type, drag, scroll, press keys, and drive UI on the user's Mac. Use when the user asks to click something, fill in a field, press a shortcut, select a menu item, or perform any action in a Mac app.

1 other project uses a version of this, 1 of them edited.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 1

    1. calebnewtonusc/Chewbacca★ 892% same

    Wrote their own version · 0

    None found.

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository10 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Refs are scoped to the snapshot that produced them: After anything that changes the UI (a sheet opens, a page navigates, a list scrolls, a …

    line 21
  2. Distinctive:

    Before any of the above, check whether the app will just do it:

    line 65
  3. Very specific:

    Always. see → act → see. Twenty steps at 95% each is 36% overall, and the failure usually happens on the step you were not watching.

    line 77

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Acting on a Mac

Click

Best to worst, in order.

nova click "@s8f3k2p9:e12"          # a ref from a snapshot. use this.
nova click "Sign In" --app Safari    # by label
nova click "412,208"                 # a raw coordinate. last resort.

A ref is unambiguous. A label is usually fine. A coordinate is a guess that happens to be right, and it stops being right when the window moves.

Refs are scoped to the snapshot that produced them. After anything that changes the UI (a sheet opens, a page navigates, a list scrolls, a window resizes), re-snapshot. A stale ref does not error, it clicks the wrong thing.

Type

nova type "hello"
nova type "$LONG_TEXT" --paste       # clipboard, auto for >200 chars

For anything long, use the clipboard. Character-by-character typing is slow and every keystroke is a chance for a dropped event. peekaboo paste sets the clipboard, sends Cmd+V, and restores what the user had copied, which matters more than it sounds.

Better than typing at all, when the app supports it: set the value directly through the tree.

agent-desktop type @s8f3k2p9:e4 "hello"

If typing silently does nothing, it is Secure Input. A password field somewhere has focus and macOS is discarding your keystrokes with no error. nova doctor --secure-input names the process holding it. Fix by moving focus off the field: open -a Finder, then retry.

Keys, menus, windows

peekaboo hotkey cmd+s
peekaboo press escape
peekaboo menu click --app Safari --path "File > New Window"
peekaboo window list --app Mail
peekaboo dialog click --button "OK"
peekaboo scroll down --amount 5
peekaboo drag --from 100,100 --to 400,400

Menu paths are more reliable than hunting for a button. If the action exists in a menu, use the menu.

Ask the app instead

Before any of the above, check whether the app will just do it:

sdef /Applications/Foo.app | head -50            # does it have a dictionary?
nova run 'tell application "Mail" to send message 1'

An AppleScript either works or returns an error. It does not have a success rate. Every action you push down from clicking to scripting leaves the probabilistic regime.

Verify

Always. see → act → see. Twenty steps at 95% each is 36% overall, and the failure usually happens on the step you were not watching.

The gate

Send, pay, delete, post, submit, reply-all, purchase, confirm: stop and ask. State exactly what you are about to do. This does not lift because the user said to act without asking; what needs authorizing is the outbound consequence, not your autonomy.

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.

Check your setup, free