web_read
Reads the current page of a web session so you can reason over it. mode='text' (visible text, default), 'a11y' (compact accessible tree of links/buttons/fields — best for deciding what to click), or 'html' (raw DOM). Returns an explicit no_session error if the session isn't open — never a silent ...
This record as markdown: /tools/com-local-mcp-local-mcp/web-read.md
What web_read does on Local
AI agents call web_read to retrieve information from Local without modifying anything. It is typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
| Parameter | Type | Required | Description |
|---|---|---|---|
mode | string | — | What to return (default text). |
session | string | — | Session name (default 'default'). |
Parameters from the server's own tool schema.
Why web_read is rated Low
This tool retrieves and queries web page content in various formats for reasoning purposes. It has no side effects, does not create, modify, delete, or execute operations. It is purely observational, making it a Read category tool with low severity since misuse would only expose information the user can already access on their own screen.
From the tool's definition Tool description explicitly states it 'Reads the current page' and 'Returns an explicit no_session error if the session isn't open — never a silent empty.' The three modes (text, a11y, html) all retrieve and return data without modifying state.
Attacks that exploit this kind of access
The rule that runs web_read safely
PolicyLayer is an MCP gateway: it sits between your AI agents and Local, and checks every tool call against a rule you set before the call runs. Nothing changes on the server itself. For web_read, this is the rule to start with:
web_read is read-only, so it stays allowed. Everything else on the server is denied unless you say otherwise.
The button opens the PolicyLayer dashboard: create your workspace, connect Local, apply this rule, and every web_read call is checked against it from then on.
Questions about web_read
Reads the current page of a web session so you can reason over it. mode='text' (visible text, default), 'a11y' (compact accessible tree of links/buttons/fields — best for deciding what to click), or 'html' (raw DOM). Returns an explicit no_session error if the session isn't open — never a silent empty. It is categorised as a Read tool in the Local MCP Server, which means it retrieves data without modifying state.
web_read accepts 2 parameters: mode, session. The full parameter table on this page comes from the server's own tool schema.
Register the Local MCP server in PolicyLayer and add a rule for web_read: 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 Local. Nothing to install.
web_read 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 web_read 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 web_read. 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.
web_read is provided by the Local MCP server (local-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
More on Local, and thousands of servers like it.
This server
Across the catalogue