desktop-automation-protocol
Skill50 starsChanged 5 months agoMITView on GitHub
How to use windows-computer-use-mcp effectively — tool selection, focus rules, safety, and mission design.
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
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 23
After automation starts, do not Alt+Tab or touch the mouse
- Distinctive: line 25
If HITL approval appears, approve it and let the target window regain focus
- Distinctive: line 26
Prefer reading logs between paused steps or on a second monitor
- Very specific: line 29
Read docs/SAFETY.md before production use
≈ “Common” is an estimate: at least 1 in 5 of the 308,232 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
Desktop automation protocol (windows-computer-use-mcp)
Tool selection guide
| Goal | Tool | Example |
| Click a known button | automation_elements(click, title="Save") | fast, needs exact title |
| Click by intent | automation_smart(click="the Save button") | slower, finds across all windows |
| Complex multi-step | automation_mission(run="save file as PDF") | autonomous loop with retry |
| Record + replay | automation_macro(record) → stop → replay | repeatable sequences |
| Wait for event | automation_watch(start, "window_appears", "Update") | event-driven triggers |
| Explore desktop | automation_smart(discover) | get all running apps |
| Extract + chain | automation_mission(workflow, actions, store_as, dollar-ref) | cross-app data pipeline |
Focus rules
- After automation starts, do not Alt+Tab or touch the mouse
- The target window must stay in foreground during click/type sequences
- If HITL approval appears, approve it and let the target window regain focus
- Prefer reading logs between paused steps or on a second monitor
Safety
- Read docs/SAFETY.md before production use
- Pair with virtualization-mcp for sandbox/VM isolation when running untrusted automation
- Opt-in features (face, keylogger) are off by default — enable only when needed
Mission design (automation_mission)
For run operation: provide a clear, concrete goal. The server decomposes via LLM sampling.
For workflow operation: provide explicit steps with tool, params, label, and optional store_as/dollar-ref for data chaining.
For self-healing: set app_path on steps so the mission can re-launch crashed apps.
In this repository5 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- .github/copilot-instructions.mdCopilot instructions
- llms-full.txtllms.txt
- llms.txtllms.txt
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.