AI agents call capture_screenshot to retrieve information from Pl without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
Screenshot capture is a read-only operation that observes the current state of the application UI. While saving to a file could be considered a write operation, the primary action is data retrieval/observation. The severity is low because screenshots cannot directly harm system state, though they may expose sensitive visual information.
From the tool's definition Tool description states 'Capture a screenshot of the current application window. Optionally save to a file.' - this retrieves visual data from the UI without modifying, executing commands, or deleting anything.
Attacks that exploit this kind of access
Capture a screenshot of the current application window. Optionally save to a file. It is categorised as a Read tool in the Pl MCP Server, which means it retrieves data without modifying state.
Register the Pl MCP server in PolicyLayer and add a rule for capture_screenshot: 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 Pl. Nothing to install.
capture_screenshot 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 capture_screenshot 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 capture_screenshot. 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.
capture_screenshot is provided by the Pl MCP server (@milaboratories/pl-mcp-server). 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 →