Robot Actions — Remote Device Control

iOS accessibility audit

ios_accessibility_audit

Read-onlyIdempotent

Run an accessibility audit on the CURRENT screen of an iOS device — Apple's own accessibility audit engine (the same one Xcode's "Audit for Accessibility" button runs), so findings match what Apple reports. Audits whatever is in the foreground right now, so navigate to the screen you care about FIRST (ios_tap_by_label / ios_navigate_url), then call this. Reports contrast failures, tap targets under 44×44pt, missing/unhelpful labels, elements the accessibility engine cannot see, Dynamic Type and clipped-text problems, and wrong traits. Each issue carries the offending element's label, type and screen-point rect — the rect centre is directly tappable with ios_tap. Pass auditTypes to narrow the run (much faster on dense screens). TIMING: the audit sees the screen as it is at that instant — running it immediately after a launch or navigation, while the UI is still animating in, under-reports (measured on device: 4 issues mid-animation vs 5 once settled). Let the screen settle first. Requires iOS 17+ (errors on older) and an active iOS automation session (auto-starts if needed).

Input Schema

TableJSON Schema

Name	Required	Description
`rect`	No	Narrow the result to a region in screen POINTS, skipping the label lookup. Takes precedence over `element`. A finding is kept when its centre falls inside this rect.
`udid`	Yes	iOS device UDID
`element`	No	Narrow the result to one element: its accessibility label, resolved on device the same way ios_tap_by_label resolves it. the platform audit engine can only audit a whole app, so this filters the findings to those inside that element's bounds — it cannot surface anything the full-screen audit did not already report. Errors if the label matches nothing, rather than silently returning the whole screen.
`auditType`	No	Convenience alias for a single-entry auditTypes, e.g. "contrast".
`auditTypes`	No	Audit types to run; omit for all of them. "contrast" = Text/background contrast below the WCAG threshold; "elementDetection" = Elements the accessibility engine cannot detect; "hitRegion" = Tap targets smaller than the 44×44pt minimum; "sufficientElementDescription" = Controls with a missing or unhelpful label; "dynamicType" = Text that does not scale with Dynamic Type; "textClipped" = Text clipped at larger content sizes; "trait" = Wrong or missing accessibility traits.

Schema Changelog

Changes observed during successful MCP inspections.

Changed1 schema field changedSep 16, 2026

changedInput schema / properties / element / description

Previous value: -"Narrow the result to one element: its accessibility label, resolved on device the same way ios_tap_by_label resolves it. XCTest can only audit a whole app, so this filters the findings to those inside that element's bounds — it cannot surface anything the full-screen audit did not already report. Errors if the label matches nothing, rather than silently returning the whole screen."New value: +"Narrow the result to one element: its accessibility label, resolved on device the same way ios_tap_by_label resolves it. the platform audit engine can only audit a whole app, so this filters the findings to those inside that element's bounds — it cannot surface anything the full-screen audit did not already report. Errors if the label matches nothing, rather than silently returning the whole screen."

AddedAug 9, 2026

TDQS

A4.4/5.0

Behavior3/5

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

Annotations already declare readOnlyHint, idempotentHint, and destructiveHint=false, so safety is covered. The description adds genuinely useful non-schema context: iOS 17+ requirement, session auto-start, and the foreground-only snapshot behavior with timing pitfalls. However the timing/animation detail is the main addition; return-shape detail beyond this is thin, so a solid 3 rather than higher.

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

Conciseness4/5

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

Front-loads what it does and the pre-condition, then uses explicit 'TIMING:' and requirements sections. It is long, but nearly every sentence carries actionable information (Xcode parity, tappable rects, iOS 17+, auto-start); a little could be trimmed but nothing is filler.

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

Completeness5/5

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

With no output schema, the description still describes the return payload (per-issue label, type, screen-point rect), the failure modes (errors on pre-iOS 17, missing session), and pre-call sequencing. An agent has everything needed to invoke and interpret it correctly.

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%, so the baseline is 3, but the description adds value: it explains that filtering by `element`/`rect` narrows findings rather than expanding the audit surface, that the rect centre is directly tappable via ios_tap, and that `auditTypes` speeds up dense screens. These are operational semantics the schema alone doesn't 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?

States a specific verb and resource (run accessibility audit on current iOS screen) and pins the exact engine (Apple's own, same as Xcode's 'Audit for Accessibility'). It also enumerates the finding categories, making it unmistakably distinct from ios_voiceover and android_accessibility_audit.

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

Usage Guidelines5/5

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

Explicit sequencing guidance: navigate to the target screen first via ios_tap_by_label / ios_navigate_url, then call this. Adds a when-not caveat about timing (running mid-animation under-reports), with concrete before/after measurements, and notes auditTypes narrows/speeds up the run.

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

All Robot Actions — Remote Device Control tools Inspect

Try in Browser

Let your AI agent see the webWeb-to-screenshot MCP for Claude, Cursor and Codex. No cookie banners. Free to start.snap-render.comAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources

Need Help?