AI agents call describe_screen to retrieve information from Uitars without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves and queries visual information from the screen (read operation). It has no capability to modify, delete, execute, or move money. The low severity reflects that screen description alone poses minimal risk—an agent gaining this capability could gather visual intelligence but cannot directly manipulate system state or trigger external actions.
From the tool's definition Tool name 'describe_screen' and description 'Describe what's visible on screen: windows, UI elements, layout' indicate data retrieval with no modifications or side effects.
Attacks that exploit this kind of access
Describe what's visible on screen: windows, UI elements, layout. It is categorised as a Read tool in the Uitars MCP Server, which means it retrieves data without modifying state.
Register the Uitars MCP server in PolicyLayer and add a rule for describe_screen: 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 Uitars. Nothing to install.
describe_screen is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the describe_screen 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 describe_screen. 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.
describe_screen is provided by the Uitars MCP server (lxsoftroxs/uitars-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →