メニュー
Use when implementing ANY Apple Intelligence or on-device AI feature. Covers Foundation Models, @Generable, LanguageModelSession, structured output, Tool protocol, iOS 26 AI integration.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Use when ANY iOS build fails, test crashes, Xcode misbehaves, or environment issue occurs before debugging code. Covers build failures, compilation errors, dependency conflicts, simulator problems, environment-first diagnostics.
Use when writing ANY async code, actors, threads, or seeing ANY concurrency error. Covers Swift 6 concurrency, @MainActor, Sendable, data races, async/await patterns.
Use when integrating ANY iOS system feature - Siri, Shortcuts, widgets, IAP, localization, privacy, alarms, calendar, reminders, contacts, background tasks, push notifications, timers. Covers App Intents, WidgetKit, StoreKit, EventKit, Contacts.
Use when working with camera, photos, audio, haptics, ShazamKit, or Now Playing. Covers AVCaptureSession, PHPicker, PhotosPicker, AVFoundation, Core Haptics, audio recognition, MediaPlayer, CarPlay, MusicKit.
Use when implementing or debugging ANY network connection, API call, or socket. Covers URLSession, Network.framework, NetworkConnection, connection diagnostics.
Use when app feels slow, memory grows, battery drains, or diagnosing ANY performance issue. Covers memory leaks, profiling, Instruments workflows, retain cycles, performance optimization.
| name | axiom-ai |
| description | Use when implementing ANY Apple Intelligence or on-device AI feature. Covers Foundation Models, @Generable, LanguageModelSession, structured output, Tool protocol, iOS 26 AI integration. |
| license | MIT |
You MUST use this skill for ANY Apple Intelligence or Foundation Models work.
Use this router when:
First, determine which kind of AI the developer needs:
| Developer Intent | Route To |
|---|---|
| On-device text generation (Apple Intelligence) | Stay here → Foundation Models skills |
| Custom ML model deployment (PyTorch, TensorFlow) — classic Core ML | See skills/ios-ml.md (hub) → conversion / compression / training files |
| Custom LLM-scale / transformer model on-device (27-cycle) | See skills/core-ai.md → Core AI conversion, runtime, specialization |
| Computer vision (image analysis, OCR, segmentation) | /skill axiom-vision → Vision framework |
| Cloud API integration (OpenAI, generic HTTP) | /skill axiom-networking → URLSession patterns |
| Cloud Claude integration (Anthropic SDK, Messages API, Claude Agent SDK) | See claude-api skill (external) → includes automated Opus 4.6 → 4.7 migration |
| System AI features (Writing Tools, Genmoji) | No custom code needed — these are system-provided |
Key boundary: Foundation Models vs ML (custom models)
When developers say "I need to train / fine-tune / personalize a model," four distinct paths exist. They are often conflated; each has different output, lifecycle, and runtime compatibility.
| Path | Trains | Output | Lifecycle | Routes to |
|---|---|---|---|---|
| FM custom adapter (26-cycle only — runtime obsoleted in 27.0) | Apple's frozen on-device 3B LLM (rank-32 LoRA) | .fmadapter package, ~160 MB | Build-time per OS version, delivered via Background Assets | skills/foundation-models-adapters.md (discipline) + skills/foundation-models-adapters-ref.md (toolkit + runtime) + skills/foundation-models-adapters-diag.md (failure modes); delivery via axiom-integration (skills/background-assets.md) |
Core ML MLUpdateTask | Your NN-spec model's fully-connected and convolutional layers | Updated .mlmodelc saved to disk | Runtime, per-user (on-device personalization) | skills/coreml-training.md |
| Create ML | A new Core ML model from scratch / transfer learning | .mlmodel | Build-time, on Mac or iOS (per type) | skills/coreml-training.md |
MLX LM (mlx_lm.lora) | Open-source LLMs on Apple silicon | adapters/adapters.safetensors — NOT loadable by Foundation Models | Build-time; not an iOS distribution path | External — outside Axiom scope; treat as adjacent research tool |
| Server LLM fine-tune | Cloud-hosted model (e.g., vendor fine-tunes) | Cloud artifact, accessed via API | Build-time; runs in cloud | /skill axiom-networking for the API integration; the fine-tune workflow is the vendor's domain |
Critical distinctions:
.safetensors) cannot be loaded into a LanguageModelSession. Different toolchain, different deployment target.MLUpdateTask is NN-spec only — does not support ML Program (.mlpackage) models from modern PyTorch / TensorFlow conversion. This is the main reason it's rarely used in new projects.skills/foundation-models.md for the deflection ladder.For the full "which path applies to me?" disambiguation (decision tree, the three week-costing mistakes, per-path routing) → skills/training-paths.md.
Foundation Models + concurrency (session blocking main thread, UI freezes):
await or running on @MainActorFoundation Models + data (@Generable decoding errors, structured output issues):
Foundation Models + security (prompt injection, securing agent tools, confirmation gating):
.onToolCall confirmation, .historyTransform spotlighting/redaction, lock-screen intent policy) → axiom-security (skills/agentic-security.md)skills/ios-ml.md is the hub (deployment, runtime, speech-to-text). The lifecycle stages have dedicated files:
skills/coreml-conversion.md (coremltools.convert, ML Program vs NN-spec, parity validation)skills/coreml-compression.md (the PTQ-vs-QAT decision, palettization/quantization/pruning)skills/coreml-training.md (Create ML; MLUpdateTask and its NN-spec-only limitation)OS27)skills/core-ai.md covers Core AI, the on-device inference framework that powers Apple Intelligence and is now open to your apps. Route here (not skills/ios-ml.md) when the model is LLM-scale / a transformer, or when the developer needs custom Metal kernels, multi-function assets, ahead-of-time compilation, KV-cache states, or the specialization/caching deployment model. Covers the Python toolchain (coreai-torch/coreai-opt), the Swift runtime (import CoreAI → AIModel/InferenceFunction/NDArray), specialization discipline, and the Foundation Models bridge (CoreAILanguageModel from the open-source coreai-models package — not a system-framework type).
Implementation patterns → skills/foundation-models.md
OS27)API reference → skills/foundation-models-ref.md
OS27: Private Cloud Compute, multimodal Attachment + ImageReference tool args, LanguageModel protocol + capabilities, reasoning + token usage, Dynamic Profiles (full modifier surface + @SessionProperty), Dynamic Instructions, custom model providers (LanguageModelExecutor), LanguageModelError migration, built-in system tools, improved Foundation Models InstrumentDiagnostics → skills/foundation-models-diag.md
Guardrails & safety decisions → skills/foundation-models-guardrails.md
permissiveContentTransformations vs .defaultMeasuring feature quality (Evaluations framework, OS27) → skills/foundation-models-evaluations-ref.md
Evaluation, Metric, Evaluator, run via Swift Testing .evaluates)ModelSample/ArrayLoader) + synthesizing more (makeSamples/SampleGenerator)ModelJudgeEvaluator, ScoringScale)ToolCallEvaluator, TrajectoryExpectation)Custom adapter training (after Approach Triage rungs 1-4) → skills/foundation-models-adapters.md
Adapter toolkit & runtime API → skills/foundation-models-adapters-ref.md
examples.train_adapter, examples.train_draft_model, examples.generate, export.export_fmadapterSystemLanguageModel.Adapter runtime API and AssetError casescom.apple.developer.foundation-model-adapter entitlementAdapter-specific diagnostics → skills/foundation-models-adapters-diag.md
compatibleAdapterNotFound, invalidAdapterName, invalidAssetcoremltools.libmilstoragepython missing on exportAutomated scanning → Launch foundation-models-auditor agent or /axiom:audit foundation-models
Detects anti-patterns AND architectural gaps:
respond(), manual JSON parsing, missing specific error catches (guardrail / contextWindow), session created per-tap, no streaming for long output, missing @Guide constraints, nested non-@Generable types, no fallback UI@Generable enums without @frozen (future-case crash), missing Cancel UX, missing transcript trimming, stale availability cache after Settings toggle, partial-output validation gaps, Tool errors indistinguishable from session errors, no retry on transient errorsScores: PRODUCTION-READY / NEEDS HARDENING / FRAGILE
coreml-conversion.md), compress (coreml-compression.md), or train/personalize (coreml-training.md). LLM-scale / transformer / 27-cycle custom model? → skills/core-ai.md (Core AI)OS27 Evaluations framework)| Thought | Reality |
|---|---|
| "Foundation Models is just LanguageModelSession" | Foundation Models has @Generable, Tool protocol, streaming, and guardrails. foundation-models covers all. |
| "I'll figure out the AI patterns as I go" | AI APIs have specific error handling and fallback requirements. foundation-models prevents runtime failures. |
| "I've used LLMs before, this is similar" | Apple's on-device models have unique constraints (guardrails, context limits). foundation-models is Apple-specific. |
| "I know the Anthropic SDK already" | Opus 4.7 removed temperature, top_p, top_k, and prefill from the Messages API. Code that worked on 4.6 returns HTTP 400 at runtime. Read claude-api (external) before changing model IDs. |
| "We need to train a custom adapter to fix the model's outputs" | Most "we need an adapter" requests resolve via rungs 1-4 of the Approach Triage (prompt engineering, @Generable/@Guide, tool calling, built-in content-tagging adapter). foundation-models has the ladder; foundation-models-adapters is only justified after each rung's failure is documented. |
| "We trained one adapter, ship it for all our users" | Each .fmadapter pins to one base-model version; one adapter does not cover a multi-OS install base. foundation-models-adapters covers per-OS variant strategy and compatibleAdapterIdentifiers(name:) runtime selection. |
| "Skip locale-specific eval, our users are mostly English-speaking" | Apple's 2025 tech report groups eval as English-US / English-outside-US / PFIGSCJK. English-only eval against a multi-locale app ships invisible non-English regressions. foundation-models-adapters covers the four-axis eval requirement. |
| "Just bundle the .fmadapter file in the app" | Apple's docs explicitly prohibit this. Adapters ship via Background Assets onDemand policy. axiom-integration (skills/background-assets.md) covers the delivery half. |
| "We'll add a custom adapter for our iOS 27 app" | The custom-adapter runtime (SystemLanguageModel.Adapter) is obsoleted in 27.0 and does not compile on a 27 deployment target — no replacement in the 27 SDK. foundation-models-adapters covers the pivot: rungs 1-4 or a custom provider (LanguageModelExecutor). |
Cloud Claude integration (claude-api skill, ships outside Axiom). Opus 4.7 removed temperature, top_p, top_k, and prefill from the Messages API — code that built successfully on 4.6 returns HTTP 400 at runtime, not compile time. The claude-api skill automates the migration (model ID swap, sampling-param removal, prefill replacement) and enforces prompt caching from day one. Skipping it costs an afternoon of production debugging when the first 400s arrive.
Apple's on-device Foundation Models and Anthropic's cloud Claude are unrelated stacks; use both in parallel when an app needs both, and treat claude-api as mandatory reading before any Claude model-ID change ships.
foundation-models:
foundation-models-diag:
User: "How do I use Apple Intelligence to generate structured data?"
→ Read: skills/foundation-models.md
User: "My AI generation is being blocked"
→ Read: skills/foundation-models-diag.md
User: "Show me @Generable examples"
→ Read: skills/foundation-models-ref.md
User: "Implement streaming AI generation"
→ Read: skills/foundation-models.md
User: "I want to add AI to my app" → First ask: Apple Intelligence (Foundation Models) or custom ML model? Route accordingly.
User: "My Foundation Models session is blocking the UI"
→ Read: skills/foundation-models.md (async patterns) + also invoke axiom-concurrency if needed
User: "Review my Foundation Models code for issues"
→ Invoke: foundation-models-auditor agent
User: "I want to run my PyTorch model on device"
→ Read: skills/ios-ml.md (classic Core ML conversion, not Foundation Models)
User: "I want to run my own LLM / SAM segmentation model on device" / "convert a PyTorch transformer with Core AI" / "my Core AI model stalls on first launch"
→ Read: skills/core-ai.md (Core AI conversion, runtime, specialization & caching)
User: "How do I train a custom adapter for our app's summarization?"
→ Read: skills/foundation-models.md (Approach Triage rungs 1-4 FIRST), then skills/foundation-models-adapters.md only if rung-1-4 failures are documented
User: "Our adapter loaded fine on iOS 26.0 but throws compatibleAdapterNotFound on 26.1"
→ Read: skills/foundation-models-adapters-diag.md (Pattern 1)
User: "What's the toolkit setup for adapter training?"
→ Read: skills/foundation-models-adapters-ref.md (Toolkit Setup)
User: "How do we ship a custom adapter to users?"
→ Read: skills/foundation-models-adapters.md (runtime lifecycle) + axiom-integration (skills/background-assets.md) (delivery)
User: "How do I measure if my prompt change made the tagging feature better?" / "Write an eval suite for my AI feature"
→ Read: skills/foundation-models-evaluations-ref.md (Evaluations framework — Metrics, Swift Testing .evaluates, model-as-judge, tool-call eval)