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 koshkinvv/ios-agent-skills --skill ios-performancegit clone --depth 1 https://github.com/koshkinvv/ios-agent-skillsWrote 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/koshkinvv/ios-agent-skills/ios-performance)<a href="https://agentmods.dev/skills/koshkinvv/ios-agent-skills/ios-performance"><img src="https://agentmods.dev/badge/skills/koshkinvv/ios-agent-skills/ios-performance/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/koshkinvv/ios-agent-skills/ios-performance"><img src="https://agentmods.dev/badge/skills/koshkinvv/ios-agent-skills/ios-performance.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 | $0.00201 | $0.03089 |
| Opus 5 | $0.00101 | $0.01545 |
| Sonnet 5 | $0.00040 | $0.00618 |
| Haiku 4.5 | $0.00020 | $0.00309 |
Grade A, and why
ios-performance 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 10d ago.
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 — 316 lines — stays where its author put it; the contents beside it link to each section on GitHub.
iOS Performance Optimization Skill
Core Rules
- Use
[weak self]in escaping closures by default — use[unowned self]only when the closure's lifetime is strictly shorter than the captured object (e.g., parent owns child, child's closure references parent). - Non-escaping closures do NOT need
[weak self]—map,filter,forEach,reduce,compactMap,sorted(by:)are non-escaping. The closure executes synchronously and releases captures immediately. - Delegates must be
weak var— the delegate protocol must conform toAnyObject(or be marked@objc). Strong delegates create retain cycles between owner and delegate. - Use
@ObservableoverObservableObject(iOS 17+) —@Observabletracks property access per-view, so only views reading a changed property re-evaluate.ObservableObjectwith@Publishedinvalidates ALL observing views on ANY published property change. - Break SwiftUI views into small subviews — each subview localizes its state dependency, so changes only re-evaluate the subview, not the entire parent body.
- Use
LazyVStack/LazyHStackinsideScrollViewfor large collections —VStackevaluates ALL children upfront. UseListfor very large datasets (10,000+) since it reuses cells. - Use
.taskmodifier instead of.onAppear+Task {}—.taskautomatically cancels when the view disappears. Use.task(id:)to restart when a dependency changes. - Profile with Instruments BEFORE optimizing — measure first, optimize second. Never guess where the bottleneck is.
- Never block the main thread — all file I/O, network calls, JSON decoding of large payloads, image processing, and database queries must run off the main thread.
- Reuse expensive objects —
URLSession,DateFormatter,JSONDecoder,NumberFormatter,NSRegularExpressionare expensive to create. Use shared instances or caches.
Performance Targets
| Metric | Target | Critical Threshold |
|---|---|---|
| Cold launch | <400ms to first frame | >2s triggers watchdog kill |
| Warm launch | <200ms | >1s feels broken |
| Frame rate | 60 FPS (120 on ProMotion) | <45 FPS is noticeable jank |
| Frame budget | 16.67ms (8.33ms at 120Hz) | >33ms = visible dropped frame |
| Memory (typical) | <100MB resident | >500MB risks jetsam on older devices |
| Memory (spike) | <200MB peak | Varies by device class |
| CPU idle | <3% | >5% drains battery |
| CPU active task | <80% sustained | 100% = thermal throttling |
| API response (perceived) | <200ms | >1s needs loading indicator |
| App size (download) | <50MB (OTA limit: 200MB) | >200MB requires Wi-Fi |
| Disk writes | <1MB/min sustained | Excessive writes degrade flash |
What ships with it
4 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 10d ago First seen · 316 lines · 201 tokens per session scan A d19340c8fe2a
ios-performance is a skill published in the GitHub repository koshkinvv/ios-agent-skills (6 stars, last pushed 5mo ago), licensed MIT. It adds 201 tokens to every session and 3,089 once invoked, about $0.0010 per session on Opus 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 2026-08-31.
Other skills, from other repositories
flutter
Use when building, structuring, testing or optimizing a Flutter app — feature-first layering, Riverpod 3 or Bloc, typed gorouter, freezed models, a dio data layer, Material 3, jank hunting, widget/golden tests. Targets Flutter 3.44 / Dart 3.12. NOT React Native (that is react-native), NOT Compose Multiplatform (that…
react-native
Use when writing the JS/TS inside a React Native or Expo app — screens, Expo Router navigation, lists, Reanimated gestures, platform forks, offline state, native modules — or killing jank and render storms. NOT eas build/submit/OTA/config-plugin (that is expo), NOT a Dart app (that is flutter), NOT web React/DOM (that…
compose-multiplatform
Use when building one shared Compose UI in Kotlin across Android, iOS, and desktop — commonMain @Composables, expect/actual, source-set placement, native interop, multiplatform ViewModel/navigation/Koin. NOT a single-platform native build (that is kotlin-android / swift-ios), and NOT Dart/Flutter cross-platform UI…
ios-accessibility
Production-grade iOS accessibility skill covering VoiceOver, Dynamic Type, color contrast, motion, Switch Control, Voice Control, and WCAG 2.2 compliance for both SwiftUI and UIKit. This skill should be used when creating new iOS screens or views, reviewing existing iOS code for accessibility, adding VoiceOver…
ios-localization
Production-grade iOS localization skill covering String Catalogs (.xcstrings), CLDR pluralization, SwiftUI/UIKit localization APIs, RTL layout, date/number/currency formatting, and enterprise patterns (modular apps, white-label, accessibility localization). This skill should be used when creating or editing localized…
ios-logging
Enterprise skill for iOS production error observability and logging (iOS 15+, Swift 5.5+). The trigger is OBSERVABILITY intent — the user wants errors captured and visible in production, not just handled. Use when: adding os.Logger or replacing print() with structured logging; setting up or integrating a crash SDK…