SKILLEMALL.ai
GitHub ↗ RU Sign in

CB remote-macos-computer-use

Set up and use cua-driver (an MCP server) so an AI agent running on a remote/cloud host can drive a macOS desktop, with a persistent background daemon + reverse-SSH-tunnel wiring that survives reboots and stays in the background. Use when you want an agent on server A to click/type/capture on a Mac on another network, or to onboard a remote desktop onto any MCP-capable agent (Hermes, Claude Code, Codex, Cursor, OpenCode). Covers install, macOS TCC grants, remote login, reverse tunnel, per-agent MCP config, health checks, and safety (bounded mode).

Not recommendedcritical or high security findings
ClawHub Agent Skills author: 东海大鲨鱼(DSBS) v0.1.0 MIT-0 7 files · 3 scripts body ≈ 1 914 tokens Open the sourceclawhub.ai↗ analyzed 2 d ago

Set up and use cua-driver (an MCP server) so an AI agent running on a remote/cloud host can drive a macOS desktop, with a persistent background daemon +…

As a process B 68/100 · Nearly there — weak spots: result and completion, progress reporting

ProcedureAI and agentstype and topics are labelled automatically from the skill text
Runs in: OpenClaw Claude / Claude Code ChatGPT / Codex Gemini CLI Cursor GitHub Copilot / VS Code Qwen Code Kimi Code CLI Z.ai (GLM) DeepSeek (Deep Code) Mistral Vibe LangChain (deepagents) Goose OpenCode
JSON
Technical rating
C
73/100
safety, quality, tests
Safety 60%
64
Quality 40%
87
Run on models
none yet
Process rating
B
68/100
Nearly there
Result and completion w 14
0
Progress reporting w 2
0
When it triggers w 12
50
the three weakest of ten parameters · all ten

What is at stake

The skill contains fragments that, in the wrong hands, cost money or data. Below: what the installer risks and what the author should do.

Dangerous commands
If you install

The skill contains commands that delete files, rewrite disks or execute code fetched from the network. The agent may run them without asking if it believes the instructions require it.

For the author

Replace destructive commands with safe equivalents that ask for confirmation, scope them to one folder, and stop piping curl into a shell: pin a version and a checksum.

Exfiltration
If you install

The instructions or scripts send data out: environment variables, keys, file contents, chat history. You may never notice, because the agent performs the upload as "part of the task".

For the author

If the upload is not needed for the task, remove it: catalogs flag such skills and delist them. If it is needed, name the destination explicitly, say what leaves the machine, and give the user a switch.

How to improve

  1. Address the high-severity findings: each costs 18 safety points. If one is a false positive, add the rule id to guard.allow in spec.yaml.
For the model run — optional
  • Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
  • A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.

Guard findings · 2

  • high Dangerous commands cmd-persistence scripts/setup-mac.sh:118
    Persistence mechanism (cron / launchd / scheduled task / autorun registry)
    launchctl load "$LA_DIR/$p.plist" && echo "    loaded $p"
  • high Exfiltration exfil-read-secret-files SKILL.md:58
    Reads credential / secret files
    ssh-copy-id -i ~/.ssh/id_e…pub -p <revport> <mac_user>@127.0.0.1   # only after tunnel is up

Files scanned: 7. Evidence is masked. Grey chips explain why severity was lowered.

Against the Agent Skills spec

✓ No remarks against the Agent Skills spec

Process rating: all ten parameters 68/100

  • 0Result and completion. Does not say what the result is
  • 0Progress reporting. Says nothing while it works
  • 50When it triggers. No condition that starts the skill
  • 60Tools and files. Uses tools (bash, web) that frontmatter does not declare
  • 70Inputs and preconditions. Inputs and preconditions are listed
  • 100Steps. 29 steps
  • 100Failures and branches. 1 branches, has a failure section
  • 100Consistency. Name and required fields are in place
  • 100Execution cost. Instruction body is 1914 tokens
  • 100Running it twice. No mutating operations
  • medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
  • low The response is described with custom markup (4 tags): a typed call is more reliable

Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.

Quality signals

  • +5Description has no quoted example phrases that should trigger the skill
  • +4Description does not say when NOT to use the skill (false activations)
  • +3Output format is not stated: the model decides each time
  • +1No license
  • +2Single-language instructions
  • +3Description length 553: enough signal without eating the budget
  • +4Structure: 15 headings
  • +3Step-by-step instructions: 29 items
  • +4Has examples (8 code blocks)
  • +3All 3 scripts are documented

Quality base 70; lint remarks subtract, signals add up to 100. Result: 87.

External checks

ClawHub: suspicious
The skill is openly built for remote Mac control, but it sets up persistent SSH and desktop access with weak default guardrails.
LLM: suspicious (high) · 30 Aug 2026