AI agents invoke scroll_until_visible to trigger actions in Phantom. 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.
This tool performs a UI automation action (scrolling) on a mobile device or emulator until a UI element is visible. It triggers external device interactions whose effects depend on arguments (direction, target element). The description is truncated ('jusqu') but the tool name and server context make clear it performs device control actions.
From the tool's definition 'scroll_until_visible' - scrolls in a direction until an element becomes visible; triggers external operations (UI interaction) on iOS simulators, Android emulators, or real devices
Attacks that exploit this kind of access
Scroll vers le bas (ou la direction choisie) jusqu. It is categorised as a Execute tool in the Phantom MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Phantom MCP server in PolicyLayer and add a rule for scroll_until_visible: 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 Phantom. Nothing to install.
scroll_until_visible 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 scroll_until_visible 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 scroll_until_visible. 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.
scroll_until_visible is provided by the Phantom MCP server (nthimpulse/phantom-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 →