PREFERRED: Always use this tool instead of calling individual action tools (click_at, type_text, press_key, launch_app, etc.) one at a time. Combine multiple steps into a single batch call — this is dramatically faster. Returns a single screenshot by default (disable with return_screenshot=false)...
AI agents invoke batch_actions to trigger actions in Macos Control. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call — builds kicked off, notifications sent, workflows started.
This tool orchestrates multiple macOS UI control actions (clicks, typing, keypresses, app launches) in a single batch. It gives an AI agent broad, chained execution capability over the entire macOS environment. Misuse could trigger any action on the system — launching apps, typing sensitive data, navigating UIs — with a blast radius spanning the entire OS.
From the tool's definition 'Combine multiple steps into a single batch call' combining tools like 'click_at, type_text, press_key, launch_app' — executes arbitrary sequences of UI actions on macOS
Documented attack patterns abuse exactly the kind of access batch_actions gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Macos Control, and nothing reaches the server without passing your rules. This is the rule we recommend for batch_actions:
{
"version": "1",
"default": "deny",
"tools": {
"batch_actions": {
"limits": [
{
"counter": "batch_actions_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}batch_actions stays usable, but rate-capped — a runaway agent can't fire it dozens of times a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
PREFERRED: Always use this tool instead of calling individual action tools (click_at, type_text, press_key, launch_app, etc.) one at a time. Combine multiple steps into a single batch call — this is dramatically faster. Returns a single screenshot by default (disable with return_screenshot=false). Only use individual tools when you need to read the result of one action before deciding the next. Stops on first error. Max 20 actions per call. Example — open Notes and write text (1 call instead of 6): [{. It is categorised as a Execute tool in the Macos Control MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Macos Control MCP server in PolicyLayer and add a rule for batch_actions: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Macos Control. Nothing to install.
batch_actions is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.
Yes. Add a rate_limit block to the batch_actions rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for batch_actions. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
batch_actions is provided by the Macos Control MCP server (peterhdd/macos-control-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Macos Control, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
22 Macos Control tools catalogued and risk-classified — across an index of 43,000+ MCP servers.