agentsclimarketplace
Sign in

Safari mcp workflow

Skill gaelic-ghost/socket/plugins/apple-dev-skills/skills/safari-mcp-workflow

Inspect a scoped live site in Safari through Safari Technology Preview's local MCP server. Use for Safari-specific DOM, console, network, screenshot, accessibility, performance, interaction, responsive, or print-media evidence.From its SKILL.md

Install
npx -y skills add gaelic-ghost/socket --skill safari-mcp-workflow

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 6 stars6 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

5.7 KB, ~1.1k tokens by cl100k_base, as published. Nobody here has run it

Safari MCP Workflow

Purpose

Use Safari MCP for evidence from a live Safari Technology Preview tab. It owns browser-runtime observation and authorized interaction, not Safari extension architecture, app-to-Safari control, Web Inspector extension design, or cross-browser certification.

When To Use

  • Use for Safari-specific rendering, DOM, console, network, screenshot, interaction, accessibility, performance, responsive, or print-media investigation.
  • Use when a local web app needs Safari evidence before a fix is proposed or accepted.
  • Use safari-extension-control-workflow for Safari Web Extension, Safari App Extension, SafariServices, or Web Inspector extension architecture.
  • Use xcode-testing-workflow for XCTest, XCUITest, simulator, or Xcode-run validation.
  • Use explore-apple-swift-docs for Apple documentation lookup that does not need a live tab.

Single-Path Workflow

  1. Scope the target before connecting: name the origin, tab, requested state, permitted interactions, and any sensitive data boundary. Do not inspect unrelated tabs.
  2. Preflight Safari Technology Preview 247 or later, its safaridriver executable, and Safari's Enable remote automation and external agents setting (shown as Allow remote automation in some builds). Read references/setup-and-privacy.md.
  3. Register Safari MCP only with explicit approval. Use the documented local safaridriver --mcp command; do not bundle a server, silently alter global MCP configuration, or claim a connection before listing its tools.
  4. Establish page state with tab, URL, loading, and get_page_content evidence before evaluating JavaScript or taking a screenshot. Use its node uid values when an interaction must target a specific element. Then collect the smallest useful combination of console, network, DOM, JavaScript, and screenshot evidence.
  5. Perform only the authorized interaction sequence; batch related reversible steps through page_interactions with a stated purpose and the relevant node uid when that makes the recorded sequence clearer. Do not submit forms, purchase, delete, change account state, or reveal secrets without a fresh user confirmation immediately before the effect.
  6. For accessibility, inspect semantic markup, labels, ARIA, keyboard reachability, and relevant visual contrast evidence. Report findings as focused checks, not a compliance certification.
  7. For performance, use bounded page metrics and network timings. State whether the evidence is a one-run observation, not a benchmark or regression guarantee.
  8. Re-check the requested outcome after each change with independent evidence. Read references/evidence-and-validation.md for the reporting contract.
  9. Return the target, Safari Technology Preview version, observed evidence, authorized actions, unresolved uncertainty, and one focused handoff.

Inputs

  • target: required origin, URL, or already selected tab.
  • goal: required bug, behavior, accessibility, performance, or visual expectation.
  • interaction_scope: optional read-only, reversible, or an explicitly named state-changing action.
  • data_boundary: optional authenticated, private, payment, or other sensitive-page restriction.
  • viewport: optional CSS-pixel viewport or media type.
  • Defaults: use read-only inspection, one scoped tab, minimal data collection, and Safari-specific claims only.

Outputs

  • status: success, handoff, or blocked.
  • evidence: selected tab/URL, Safari Technology Preview version, tool results, screenshots when useful, and observed-versus-expected behavior.
  • actions: every interaction performed and whether it changed state.
  • next_step: one concrete fix, retest, or workflow handoff.

Guards and Stop Conditions

  • Do not expose or log credentials, cookies, tokens, AutoFill data, private text, or unrelated browsing activity.
  • Do not treat Safari MCP as a Safari extension API, an arbitrary native-control surface, or cross-browser proof.
  • Do not present a screenshot alone as behavioral proof; pair it with DOM, state, console, network, or interaction evidence as appropriate.
  • Stop with blocked when Safari Technology Preview, the driver, the required Safari setting, or an authorized target is unavailable.
  • Stop and ask immediately before any irreversible or account-, payment-, privacy-, or publication-affecting action.

Fallbacks and Handoffs

  • Recommend safari-extension-control-workflow for Safari integration and extension decisions.
  • Recommend xcode-testing-workflow for Xcode-owned tests and UI automation.
  • Recommend apple-ui-accessibility-workflow for Apple native UI accessibility work.
  • Recommend explore-apple-swift-docs for current Apple or WebKit documentation.

Customization

Use references/customization-flow.md. The first version has no runtime-enforced settings: origin, interaction, and privacy boundaries must be chosen for each live session.

References

Workflow References

  • references/setup-and-privacy.md
  • references/evidence-and-validation.md
  • references/customization-flow.md

Authoritative Sources

Script Inventory

  • scripts/customization_config.py

What ships with it: 6 files

9.9 KB alongside SKILL.md, 1 of them executable

agents/

references/

scripts/

Keep looking

Skills are one crate of 326,834. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.