Skip to main content
Glama
achiya-automation

Safari MCP Server

by achiya-automation
OverviewSchemaSourceRelated ServersDiscussionsAdmin
JavaScript
Local

safari_run_script

Runs a batch of Safari browser automation actions in one MCP session, covering tabs, navigation, reading, clicking, forms, scrolling, and evaluation.

Instructions

Batch Safari actions in one MCP session. Named profiles use the verified extension and support: newTab, switchTab, getReceipt, listTabs, closeTab, navigate, navigateAndRead, readPage, snapshot, getElementInfo, querySelectorAll, waitFor, waitForTime, click, clickAndOpenPopup, fill, fillForm, clearField, typeText, selectOption, pressKey, scroll, scrollTo, scrollToElement, hover, evaluate, reload, goBack, goForward. switchTab accepts an index or an opaque receipt; a valid receipt recovers the exact owned tab across Safari windows without focusing one. clickAndOpenPopup takes exactly one selector or snapshot ref, targets one exact frame, performs one click, captures a blocked HTTP(S) window.open, and opens it as a background tab without focusing Safari; it refuses CAPTCHA/challenge targets and never returns URL query/hash data. Non-profile mode retains the legacy action set plus getReceipt; there, closeTab and switchTab take a tab index, not a receipt, and getReceipt needs the tab's receipt in args. After a cross-origin redirect, rotate the tab's receipt with {"action":"getReceipt","args":{"receipt":""}}.

Input Schema

NameRequiredDescriptionDefault
stepsYesArray of steps to execute sequentially

Schema Changelog

Changes observed during successful MCP inspections.

  1. Addedv2.11.6
  2. Removedv2.11.3
  3. Addedv2.10.9
  4. Removedv2.10.8
  5. Addedv2.10.5
  6. Removedv2.10.4
  7. First observed

TDQS

A4.2/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries the full burden and does real work: clickAndOpenPopup refuses CAPTCHA/challenge targets, never returns URL query/hash data, and opens background tabs without focusing Safari. Receipt ownership semantics and the cross-origin receipt-rotation requirement are disclosed. Missing: what happens when a step in the batch fails (abort vs continue) and what the call returns.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness3/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Purpose is front-loaded in sentence one, but the body is a dense semicolon-chained wall containing a 30-item action enumeration and three distinct behavioral topics with no grouping or line breaks. Most of the content earns its place, yet it is harder to scan than it needs to be.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

No output schema and no annotations, so the description must cover the surface itself, and it largely does: action vocabulary, profile vs legacy mode differences, receipt lifecycle, and popup edge cases. The main remaining gap is batch failure/return semantics, which an agent needs before committing a multi-step script.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100% but the schema only defines `steps` as an array of {action, args} with a free-form action string and a one-line example. The description supplies the actual action vocabulary (~30 names) and the argument conventions for switchTab/closeTab/getReceipt per mode, which is meaning the schema cannot convey.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

Opens with a specific verb+resource+scope: batch Safari actions in one MCP session. That scope (many actions in one call) is exactly what separates it from the ~70 single-action siblings like safari_click or safari_navigate, so an agent can route to it without opening a schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives clear operating context: named profiles use the verified extension and expose one action set, while non-profile mode retains a legacy set with different closeTab/switchTab argument types. It also prescribes a follow-up after cross-origin redirects (rotate the receipt via getReceipt). It stops short of stating when to prefer batching over the individual sibling tools.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Deploy Server

Other Tools