Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add cvs-health/ios-swiftui-accessibility-techniques --skill figma-annotationsgit clone --depth 1 https://github.com/cvs-health/ios-swiftui-accessibility-techniquesWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/cvs-health/ios-swiftui-accessibility-techniques/figma-annotations)<a href="https://agentmods.dev/skills/cvs-health/ios-swiftui-accessibility-techniques/figma-annotations"><img src="https://agentmods.dev/badge/skills/cvs-health/ios-swiftui-accessibility-techniques/figma-annotations/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/cvs-health/ios-swiftui-accessibility-techniques/figma-annotations"><img src="https://agentmods.dev/badge/skills/cvs-health/ios-swiftui-accessibility-techniques/figma-annotations.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00110 | $0.21976 |
| Opus 5.5 | $0.00044 | $0.08790 |
| Sonnet 5.5 | $0.00022 | $0.04395 |
| Haiku 4.5 | $0.00011 | $0.02198 |
Grade A, and why
figma-a11y-annotations scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured today.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 914 lines — stays where its author put it; the contents beside it link to each section on GitHub.
iOS Accessibility Design Annotations
You are annotating a native iOS design so that an engineer can build it accessibly in SwiftUI. These rules are mandatory and override your defaults.
An annotation is a specification, not a report. It exists for exactly one reason: to tell the developer something they cannot work out from looking at the design. If a developer would build it correctly without the note, the note is noise — and noise is the main way AI annotation fails. An over-annotated frame is worse than an un-annotated one, because the reviewer stops reading and the three annotations that mattered get lost with the forty that didn't.
Read the principles, "Rule 0", and "The gate" before writing a single annotation.
Principles
Ten ideas the rest of this file is built on. When a specific rule below is unclear, resolve it against these.
1. Never annotate visible text as an accessibility label
If a control's visible text is already its accessible name, write no Name: annotation at all. This is the most common and most damaging annotation mistake, and it is damaging in three separate ways:
- It clutters. A button labeled "Refill" annotated
Name: Refillhas spent a note to say nothing. Do this across a screen and the handful of notes that carry real information are buried. - It introduces drift, and manufactures a failure that did not exist. An annotated label is retyped prose, so it is rarely byte-identical to the visible text — "Refill" becomes "Refill prescription", "Add to Cart" becomes "Add to cart", "Done" becomes "Save". The moment it differs, the accessible name no longer contains the visible label, and you have created a WCAG 2.5.3 Label in Name failure — along with a control that Voice Control users cannot activate by saying what they see. The code was correct before you annotated it.
- It produces redundant code. A developer handed
Name: Refillwill write.accessibilityLabel("Refill")on aButton("Refill"), because the annotation implies something was needed. Now the same string exists twice, must be localized twice, and will be updated in one place and not the other. SwiftUI already derived that name from the visible text, for free and correctly.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- today Changed · +2 lines 3687cd17c137
- yesterday Changed · +248 lines · +6 tokens per session a0ceb93bc7ea
- 3d ago Changed · +1 tokens per session f00e7a23fd51
- 6d ago First seen · 664 lines · 103 tokens per session scan A b9b9c8483feb
figma-a11y-annotations is a skill published in the GitHub repository cvs-health/ios-swiftui-accessibility-techniques (384 stars, last pushed today), licensed Apache-2.0. It adds 110 tokens to every session and 21,976 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-10.
Other skills, from other repositories
liquid-glass
Implement, migrate, review, and diagnose iOS Liquid Glass interfaces in SwiftUI and UIKit. Use for system chrome, custom glass effects, transitions, accessibility, and version-specific issues in native iOS apps.
motion-system
Give an iOS app one consistent motion language: a small set of named timings, springs and haptics in a single Motion.swift, derived from the app's own code and from real shipped apps with the same feel. Use when asked to make an app's animations consistent, create motion tokens or a motion system, clean up scattered…
ttb-skill-native-swiftui-components
Build reusable native SwiftUI UI components using 100% standard SwiftUI primitives with TTBaseUIKit design tokens, localization, chainable modifiers, and project architecture rules. NO TTBaseSUI wrappers. iOS 14+.
apple-design
Complete Apple Human Interface Guidelines (HIG) and Apple Design System standard. Use when designing, building, or auditing UI/UX for iOS, iPadOS, macOS, watchOS, visionOS, or Apple-styled web and mobile applications.
feedback-states
Design how one feature or screen in an iOS app behaves while it works: loading, showing saved data while refreshing, success, empty, error and retry, with the motion, haptics and words for each state, backed by real shipped apps. Use when asked to add or improve loading states, success or error feedback, empty states…
find-animation-opportunities
Find the places in an iOS app's UI where motion would genuinely help, what to leave still, and what existing motion is broken or overdone, and recommend real shipped interactions to copy for each. Use when asked where an app should animate, how to make it feel more alive or polished, to audit or review its motion or…