macos-use
Skill396 starsChanged 6 months agoMITView on GitHub
GUI control for macOS apps via mediar-ai's mcp-server-macos-use. Click, type, scroll, key-press, open apps — driven by accessibility tree, works in non-interactive Claude Code mode. Use this for any Sutando task that needs to drive another macOS application (Safari, Zoom, Mail, Finder, etc.).
In this repository73 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- agent-activitySkill
- agent-registrySkill
- agent-room-opsSkill
- answer-formatSkill
- audio-transcribeSkill
- bee-actionsSkill
- bot2bot-postSkill
- call-diagnosticsSkill
- claude-codexSkill
- claude-geminiSkill
- claude-piSkill
- claude-routerSkill
- collaboration-intelligenceSkill
- connect-appsSkill
- context-dropSkill
- context-reconstructSkill
- db-browserSkill
- deal-finderSkill
- doc-ingestSkill
- electron-overlay-dimmingSkill
- email-findSkill
- engine-conflict-resolveSkill
- gemini-ttsSkill
- image-generationSkill
- import-claude-contextSkill
- local-workspace-serverSkill
- macos-toolsSkill
- make-viral-videoSkill
- marketplaceSkill
- media-transcriptSkill
- meeting-prepSkill
- meeting-schedulerSkill
- mention-gateSkill
- model-switchSkill
- morning-briefingSkill
- ms365-connectSkill
- observeSkill
- obsidian-vaultSkill
- openai-ttsSkill
- open-sutando-refSkill
- overlay-appsSkill
- phone-conversationSkill
- proactive-loopSkill
- pr-shepherd-contractSkill
- quota-trackerSkill
- regression-searchSkill
- relaySkill
- releaseSkill
- report-feedbackSkill
- review-preflightSkill
- room-collabSkill
- schedule-cronsSkill
- screen-companionSkill
- screen-recordSkill
- self-diagnoseSkill
- self-upgradeSkill
- session-recapSkill
- skill-usage-reportSkill
- The first 60 shown.
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
- Distinctive: line 19
Prefer this skill over:
- Very specific: line 27
After install, these appear as mcp__macos-use__* in Claude Code:
- Distinctive: line 38
Every tool returns an accessibility-tree snapshot of the target app — structured UI elements with roles, titles, positions, and identifiers.
≈ “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.
Files
macos-use
Drive macOS applications from Claude Code via mediar-ai's mcp-server-macos-use. A Swift MCP server that wraps the macOS Accessibility API. Unlike Claude's built-in computer-use, this works in non-interactive mode (which is how Sutando's proactive loop and task bridge run), does not hold a machine-wide lock, and does not require a Pro/Max subscription.
When to use
- "Open Safari and navigate to github.com" — anything requiring real GUI interaction with an app
- "Click the Join button on the Zoom invite dialog"
- "Type this into the Discord message box"
- "Scroll the frontmost window to the bottom"
- Any task that currently falls back to AppleScript + Quartz mouse events in
src/inline-tools.ts
Prefer this skill over:
bash src/screen-capture.sh— that captures screenshots;macos-useactually interacts- AppleScript
tell applicationblocks — more reliable, better error handling cliclick— lower-level, no accessibility context- Claude's built-in
computer-use— that mode requires interactive sessions and holds a lock that contends with Sutando's own loop
Tools exposed
After install, these appear as mcp__macos-use__* in Claude Code:
| Tool | Parameters | Purpose |
open_application_and_traverse | identifier (name/bundle ID/path) | Launch or activate an app, return its a11y tree |
click_and_traverse | pid, x, y | Click at coordinates in a target app, return updated tree |
type_and_traverse | pid, text | Type into the frontmost element |
press_key_and_traverse | pid, key | Press a named key (Return, Tab, Escape, arrows, ...) |
scroll_and_traverse | pid, direction, amount | Scroll in a direction |
refresh_traversal | pid | Re-read the a11y tree without acting |
Every tool returns an accessibility-tree snapshot of the target app — structured UI elements with roles, titles, positions, and identifiers. No pixels. Model reasons over the tree, not over screenshots.
Install
Two steps, one-time:
# 1. Build the Swift binary (~35s)
bash skills/macos-use/scripts/build.sh
# 2. Register with Claude Code's MCP config (writes ~/.claude.json)
bash skills/macos-use/scripts/install-mcp.sh
# 3. Grant Accessibility permission
# System Settings → Privacy & Security → Accessibility
# Click +, navigate to ~/.macos-use-mcp/.build/release/mcp-server-macos-use, enable.Restart Claude Code after install for the MCP tools to appear.
Gotchas
- Swift 6 build fragility: the
swift-sdktransitive dep has data-race errors that Swift 6.3+ strict-concurrency trips on.build.shuses-Xswiftc -swift-version -Xswiftc 5as a workaround. When upstream fixes this, remove the flag. - Accessibility permission: the binary must be added to System Settings → Privacy & Security → Accessibility, or every tool call will return "not authorized". First-run error is obvious; owner must click through once.
- Apps without good a11y trees: Canvas / Electron / games degrade badly. For those, fall back to
screen-capture.sh+ Claude vision. - Build dep pulled from GitHub: air-gapped Macs won't work. No prebuilt releases yet.
- Multi-node: each node builds its own binary. Not synced via
sutando-memory.git(binaries are machine-specific). Runbuild.sh+install-mcp.shon Mac Mini and MacBook separately.
Quick self-test
After install + restart:
Sutando, open Safari and navigate to https://github.com/sonichi/sutandoYou should see Claude invoke mcp__macos-use__open_application_and_traverse with identifier: "Safari", then type_and_traverse into the URL bar, then press_key_and_traverse with Return.
Related
- Research + decision memo:
notes/issue-65-computer-use-research.md - Issue: #65 Add Claude Computer Use support
- Upstream: https://github.com/mediar-ai/mcp-server-macos-use
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
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.

