Return JSON of visible interactive/form controls with ids (e0, e1, …).
AI agents call inspect_elements to retrieve information from Visual Annotation MCP without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool inspects and lists DOM elements, which is a read-only operation analogous to querying or fetching data. It has no capability to modify, execute, or delete anything. The blast radius of misuse is minimal—at worst, an agent learns about page structure to make better-informed decisions for subsequent actions.
From the tool's definition Tool description states 'Return JSON of visible interactive/form controls' - a retrieval operation with no modification or execution. The name 'inspect_elements' and function of returning element metadata confirm data querying without side effects.
Attacks that exploit this kind of access
Return JSON of visible interactive/form controls with ids (e0, e1, …). It is categorised as a Read tool in the Visual Annotation MCP MCP Server, which means it retrieves data without modifying state.
Register the Visual Annotation MCP server in PolicyLayer and add a rule for inspect_elements: 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 Visual Annotation MCP. Nothing to install.
inspect_elements 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 inspect_elements 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 inspect_elements. 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.
inspect_elements is provided by the Visual Annotation MCP server (mstocker1/visual_annotation_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 →