Click at screen coordinates. button: 1=left, 2=middle, 3=right.
AI agents invoke mouse_click to trigger actions in GNOME Desktop MCP. 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.
Mouse click injection is an Execute-category action: it triggers arbitrary UI interactions on the desktop whose effects depend entirely on what is at the target coordinates. An AI agent could click buttons, confirm dialogs, launch applications, or activate destructive actions. The blast radius is high because the agent can interact with any visible UI element across the entire desktop session.
From the tool's definition 'Click at screen coordinates' with button options (left/middle/right) — injects mouse input events into the live desktop session
Documented attack patterns abuse exactly the kind of access mouse_click gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and GNOME Desktop MCP, and nothing reaches the server without passing your rules. This is the rule we recommend for mouse_click:
{
"version": "1",
"default": "deny",
"tools": {
"mouse_click": {
"limits": [
{
"counter": "mouse_click_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
} mouse_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.
Free to start. No card required.
Click at screen coordinates. button: 1=left, 2=middle, 3=right. It is categorised as a Execute tool in the GNOME Desktop MCP MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the GNOME Desktop MCP server in PolicyLayer and add a rule for mouse_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 GNOME Desktop MCP. Nothing to install.
mouse_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 mouse_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 mouse_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.
mouse_click is provided by the GNOME Desktop MCP server (sbuysse/gnome-desktop-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from GNOME Desktop MCP, 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.
30 GNOME Desktop MCP tools catalogued and risk-classified — across an index of 43,000+ MCP servers.