AI agents call get_screen_text to retrieve information from Poof without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves screen output and terminal state information. While it is read-only (no side effects on data), the severity is elevated to 'medium' because the terminal context could contain sensitive information (credentials, API keys, personal data), and the broader server context (controlling macOS Terminal.app) creates a high-impact environment. However, the tool itself performs only data retrieval.
From the tool's definition Tool name is 'get_screen_text' and description states 'Get screen as plain text.' The verb 'Get' and passive retrieval framing indicate data retrieval without modification.
Attacks that exploit this kind of access
Get screen as plain text. Opens the terminal if not already open. It is categorised as a Read tool in the Poof MCP Server, which means it retrieves data without modifying state.
Register the Poof MCP server in PolicyLayer and add a rule for get_screen_text: 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 Poof. Nothing to install.
get_screen_text 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 get_screen_text 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 get_screen_text. 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.
get_screen_text is provided by the Poof MCP server (mattapperson/poof-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 →