ui_click
Clicks an element (by element_ref, at its center) or a screen coordinate (by coords). button left|right, count 2 = double-click. Returns {clicked, at:{x,y}}. Requires Accessibility permission.
This record as markdown: /tools/com-local-mcp-local-mcp/ui-click.md
What ui_click does on Local
AI agents invoke ui_click to trigger actions in Local. 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.
| Parameter | Type | Required | Description |
|---|---|---|---|
by | string | — | Default element if element_ref given, else coords. |
count | integer | — | 1 (default) or 2 for double-click. |
button | string | — | Default left. |
coords | object | — | {x,y} in global screen points. |
element_ref | string | — |
Parameters from the server's own tool schema.
Why ui_click is rated High
This tool performs UI interaction by clicking arbitrary screen elements or coordinates, which can trigger any application action on the user's Mac. An AI agent could misuse this to click buttons, confirm dialogs, or activate dangerous controls across any visible UI, making it a general-purpose Execute tool with high blast radius.
From the tool's definition Clicks an element (by element_ref, at its center) or a screen coordinate (by coords). button left|right, count 2 = double-click.
Attacks that exploit this kind of access
The rule that runs ui_click safely
PolicyLayer is an MCP gateway: it sits between your AI agents and Local, and checks every tool call against a rule you set before the call runs. Nothing changes on the server itself. For ui_click, this is the rule to start with:
ui_click 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.
The button opens the PolicyLayer dashboard: create your workspace, connect Local, apply this rule, and every ui_click call is checked against it from then on.
Questions about ui_click
Clicks an element (by element_ref, at its center) or a screen coordinate (by coords). button left|right, count 2 = double-click. Returns {clicked, at:{x,y}}. Requires Accessibility permission. It is categorised as a Execute tool in the Local MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
ui_click accepts 5 parameters: by, count, button, coords, element_ref. The full parameter table on this page comes from the server's own tool schema.
Register the Local MCP server in PolicyLayer and add a rule for ui_click: 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 Local. Nothing to install.
ui_click 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 ui_click 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 ui_click. 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.
ui_click is provided by the Local MCP server (local-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
More on Local, and thousands of servers like it.
Across the catalogue