foundation-models

Name: foundation-models
Author: johnrogers

222stars19forks

How to install foundation-models?

johnrogers/claude-swift-engineering/foundation-models

$npx -y skills add johnrogers/claude-swift-engineering --skill foundation-models

Installs into the current project.

Summary

Use when implementing on-device AI with Apple's Foundation Models framework (iOS 26+), building summarization/extraction/classification features, or using @Generable for type-safe structured output.

Files · 1

View on GitHub

SKILL.md

1	# Foundation Models
2
3	Apple's on-device AI framework providing access to a 3B parameter language model for summarization, extraction, classification, and content generation. Runs entirely on-device with no network required.
4
5	## Overview
6
7	Foundation Models enable intelligent text processing directly on device without server round-trips, user data sharing, or network dependencies. The core principle: leverage on-device AI for specific, contained tasks (not for general knowledge).
8
9	## Reference Loading Guide
10
11	ALWAYS load reference files if there is even a small chance the content may be required. It's better to have the context than to miss a pattern or make a mistake.
12
13	\| Reference \| Load When \|
14	\|-----------\|-----------\|
15	\| [Getting Started](references/getting-started.md) \| Setting up LanguageModelSession, checking availability, basic prompts \|
16	\| [Structured Output](references/structured-output.md) \| Using `@Generable` for type-safe responses, `@Guide` constraints \|
17	\| [Tool Calling](references/tool-calling.md) \| Integrating external data (weather, contacts, MapKit) via Tool protocol \|
18	\| [Streaming](references/streaming.md) \| AsyncSequence for progressive UI updates, PartiallyGenerated types \|
19	\| [Troubleshooting](references/troubleshooting.md) \| Context overflow, guardrails, errors, anti-patterns \|
20
21	## Core Workflow
22
23	1. Check availability with `SystemLanguageModel.default.availability`
24	2. Create `LanguageModelSession` with optional instructions
25	3. Choose output type: plain String or @Generable struct
26	4. Use streaming for long generations (>1 second)
27	5. Handle errors: context overflow, guardrails, unsupported language
28
29	## Model Capabilities
30
31	\| Use Case \| Foundation Models? \| Alternative \|
32	\|----------\|-------------------\|-------------\|
33	\| Summarization \| Yes \| - \|
34	\| Extraction (key info) \| Yes \| - \|
35	\| Classification \| Yes \| - \|
36	\| Content tagging \| Yes (built-in adapter) \| - \|
37	\| World knowledge \| No \| ChatGPT, Claude, Gemini \|
38	\| Complex reasoning \| No \| Server LLMs \|
39
40	## Platform Requirements
41
42	- iOS 26+, macOS 26+, iPadOS 26+, visionOS 26+
43	- Apple Intelligence-enabled device (iPhone 15 Pro+, M1+ iPad/Mac)
44	- User opted into Apple Intelligence
45
46	## Common Mistakes
47
48	1. Using Foundation Models for world knowledge — The 3B model is trained for on-device tasks only. It won't know current events, specific facts, or "who is X". Use ChatGPT/Claude for that. Keep prompts to: summarizing user's own content, extracting info, classifying text.
49
50	2. Blocking the main thread — LanguageModelSession calls must run on a background thread or async context. Blocking the main thread locks UI. Always use `Task { }` or background queue.
51
52	3. Ignoring context overflow — The model has finite context. If the user pastes a 50KB document, it will fail silently or truncate. Check input length and trim/truncate proactively.
53
54	4. Forgetting to check availability — Not all devices support Foundation Models. Check `SystemLanguageModel.default.availability` before using. Graceful degradation is required.
55
56	5. Ignoring guardrails — The model won't answer harmful queries. Instead of fighting it, design prompts that respect safety guidelines. Rephrasing requests usually works.

People who liked this also viewed

6 picks

Type

foundation-models

222stars19forks

View on GitHub

How to install foundation-models?

johnrogers/claude-swift-engineering/foundation-models

$npx -y skills add johnrogers/claude-swift-engineering --skill foundation-models

Installs into the current project.

Summary

Use when implementing on-device AI with Apple's Foundation Models framework (iOS 26+), building summarization/extraction/classification features, or using @Generable for type-safe structured output.

Files · 1

View on GitHub

SKILL.md

1	# Foundation Models
2
3	Apple's on-device AI framework providing access to a 3B parameter language model for summarization, extraction, classification, and content generation. Runs entirely on-device with no network required.
4
5	## Overview
6
7	Foundation Models enable intelligent text processing directly on device without server round-trips, user data sharing, or network dependencies. The core principle: leverage on-device AI for specific, contained tasks (not for general knowledge).
8
9	## Reference Loading Guide
10
11	ALWAYS load reference files if there is even a small chance the content may be required. It's better to have the context than to miss a pattern or make a mistake.
12
13	\| Reference \| Load When \|
14	\|-----------\|-----------\|
15	\| [Getting Started](references/getting-started.md) \| Setting up LanguageModelSession, checking availability, basic prompts \|
16	\| [Structured Output](references/structured-output.md) \| Using `@Generable` for type-safe responses, `@Guide` constraints \|
17	\| [Tool Calling](references/tool-calling.md) \| Integrating external data (weather, contacts, MapKit) via Tool protocol \|
18	\| [Streaming](references/streaming.md) \| AsyncSequence for progressive UI updates, PartiallyGenerated types \|
19	\| [Troubleshooting](references/troubleshooting.md) \| Context overflow, guardrails, errors, anti-patterns \|
20
21	## Core Workflow
22
23	1. Check availability with `SystemLanguageModel.default.availability`
24	2. Create `LanguageModelSession` with optional instructions
25	3. Choose output type: plain String or @Generable struct
26	4. Use streaming for long generations (>1 second)
27	5. Handle errors: context overflow, guardrails, unsupported language
28
29	## Model Capabilities
30
31	\| Use Case \| Foundation Models? \| Alternative \|
32	\|----------\|-------------------\|-------------\|
33	\| Summarization \| Yes \| - \|
34	\| Extraction (key info) \| Yes \| - \|
35	\| Classification \| Yes \| - \|
36	\| Content tagging \| Yes (built-in adapter) \| - \|
37	\| World knowledge \| No \| ChatGPT, Claude, Gemini \|
38	\| Complex reasoning \| No \| Server LLMs \|
39
40	## Platform Requirements
41
42	- iOS 26+, macOS 26+, iPadOS 26+, visionOS 26+
43	- Apple Intelligence-enabled device (iPhone 15 Pro+, M1+ iPad/Mac)
44	- User opted into Apple Intelligence
45
46	## Common Mistakes
47
48	1. Using Foundation Models for world knowledge — The 3B model is trained for on-device tasks only. It won't know current events, specific facts, or "who is X". Use ChatGPT/Claude for that. Keep prompts to: summarizing user's own content, extracting info, classifying text.
49
50	2. Blocking the main thread — LanguageModelSession calls must run on a background thread or async context. Blocking the main thread locks UI. Always use `Task { }` or background queue.
51
52	3. Ignoring context overflow — The model has finite context. If the user pastes a 50KB document, it will fail silently or truncate. Check input length and trim/truncate proactively.
53
54	4. Forgetting to check availability — Not all devices support Foundation Models. Check `SystemLanguageModel.default.availability` before using. Graceful degradation is required.
55
56	5. Ignoring guardrails — The model won't answer harmful queries. Instead of fighting it, design prompts that respect safety guidelines. Rephrasing requests usually works.

People who liked this also viewed

6 picks

Type