AI agents call comet_screenshot to retrieve information from Comet without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
Screenshots retrieve data (visual content) from a webpage with no side effects. There is no creation, modification, deletion, code execution, or financial impact. This clearly falls under the Read category, which includes data retrieval operations. The severity is low because viewing page content poses minimal risk, though the tool operates within a browser context managed by the parent tool set.
From the tool's definition Tool name 'comet_screenshot' and description 'Capture a screenshot of current page' indicate this retrieves visual data from a webpage without modifying, deleting, or executing code.
Attacks that exploit this kind of access
Capture a screenshot of current page. It is categorised as a Read tool in the Comet MCP Server, which means it retrieves data without modifying state.
Register the Comet MCP server in PolicyLayer and add a rule for comet_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 Comet. Nothing to install.
comet_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 comet_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 comet_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.
comet_screenshot is provided by the Comet MCP server (simplicianokelly52/comet_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 →