New Your team’s decisions, in one playbook every coding agent works from. Never answer your agent twice

browser.eval

Evaluates a JavaScript expression in the inspected page (MAIN world) and returns the JSON-serialisable result. An expression that yields nothing returns null; one that throws, or yields a value that cannot be serialised (a DOM node, a function, a circular object), is a PAGE_ERROR naming the reaso...

SERVERPixi Inspector SOURCEpixi-inspector-mcp
Low RISK CLASS
Category Other
Parameters 41 required
Recommended Rate-limitedsee the rule below
Registry record Grade F, identity unverified Pull the record →

This record as markdown: /tools/pixi-inspector-mcp/browser.eval.md

What browser.eval does on Pixi Inspector

AI agents call browser.eval as a supporting operation in Pixi Inspector workflows.

ParameterTypeRequiredDescription
tabId integer — Legacy: a bare tab id, resolved inside the browser this session is already working with. Prefer `target`, which is unambiguous across browsers.
target string — Drive a specific tab: a target id like "b7a3c1#487" from status/use_target/CHOOSE_TARGET. Omit to use this session's chosen tab (or the only available one). Pas
frameId integer — Escape hatch: drive a specific FRAME of the tab (0 is the top frame). Normally unnecessary — the extension picks the frame running the game (canvas owner first,
expression string Yes JavaScript expression to evaluate

Parameters from the server's own tool schema.

Why browser.eval is rated Low

Without any rule in front of it, every call an agent makes to browser.eval happens silently: no trail, no cap, no way to know it misbehaved until something breaks downstream.

Risk signalsAccepts freeform code/query input (expression)

Questions about browser.eval

What does the browser.eval tool do? +

Evaluates a JavaScript expression in the inspected page (MAIN world) and returns the JSON-serialisable result. An expression that yields nothing returns null; one that throws, or yields a value that cannot be serialised (a DOM node, a function, a circular object), is a PAGE_ERROR naming the reason — so a null result always means the page really produced null. Return a projection of what you need rather than a live object. Runs in the TOP frame by default — unlike the pixi.* tools, this one does not rank frames. On a multi-frame game pass frameId (from a pixi.* result's target line) to evaluate inside the frame that owns the game; the result's target line reports the frame it ran in. It is categorised as a Other tool in the Pixi Inspector MCP Server, which means it performs auxiliary operations.

What parameters does browser.eval accept? +

browser.eval accepts 4 parameters: tabId, target, frameId, expression. Required: expression. The full parameter table on this page comes from the server's own tool schema.

How do I enforce a policy on browser.eval? +

Register the Pixi Inspector MCP server in PolicyLayer and add a rule for browser.eval: 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 Pixi Inspector. Nothing to install.

What risk level is browser.eval? +

browser.eval is a Other tool with low risk. Read-only tools are generally safe to allow by default.

Can I rate-limit browser.eval? +

Yes. Add a rate_limit block to the browser.eval 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.

How do I block browser.eval completely? +

Set action: deny in the PolicyLayer policy for browser.eval. 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.

What MCP server provides browser.eval? +

browser.eval is provided by the Pixi Inspector MCP server (pixi-inspector-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.

More on Pixi Inspector, and thousands of servers like it.

Across the catalogue

// THE MCP REGISTRY

PolicyLayer tracks 44,603 MCP servers and 515,000+ tools.

Every server has a live record: who publishes it, whether it answers without auth, its risk grade, every tool classified, the recommended policy. This page is one line of Pixi Inspector's. Pull the full record:

Teams ship this data inside their own products. See what a licence covers →

// GET IN TOUCH

Have a question or want to learn more? Send us a message.

Message sent.

We'll get back to you soon.