claude-skills-collection
AGENTS.md18 starsChanged 56 days agoMITView on GitHub
This file contains shared repository guidance for any coding agent or contributor. Host-specific entry points refer here; the skill content does not depend on a particular agent. This repository is a skills collection focused on Swift and iOS development. Each skill is a self-contained directory with a SKILL.md, a user-facing README.md, and supporting examples or reference material. When adding a new skill, follow the same pattern used by existing skills:
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,774 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository14 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- app-intents-widgetsSkill
- Cross-Platform App Development SkillSkill
- iOS Accessibility SkillSkill
- iOS Animation Graphics SkillSkill
- iphone-duo-designSkill
- Memory Leak Diagnosis SkillSkill
- swift-concurrency-migrationSkill
- swiftdata-core-data-migrationsSkill
- Swift Modern Architecture SkillSkill
- Swift Performance Optimization SkillSkill
- SpeechAnalyzer Framework ExpertSkill
- SwiftUI Programming SkillSkill
- Swift Unit Testing SkillSkill
The file
Show its rules and text
We split this AGENTS.md into the rules it gives an agent
17 rules. About 0 are common, 8 are distinctive, and 9 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈8
Distinctive · estimate
This author's own choices. Worth a look.
9
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 27
Keep core instructions independent of host tool names, slash commands, plugins, and machine-specific paths.
- Distinctive: line 28
Keep skill changes scoped: update only the relevant skill folder unless a cross-cutting fix is required.
- Distinctive: line 30
Prefer modern Swift patterns in examples (Swift Concurrency, Observation, SwiftData when appropriate).
- Distinctive: line 31
Verify code snippets remain coherent and compile-ready when copied into a project.
- Distinctive: line 33
Use lowercase kebab-case for new skill directory names.
- Very specific: line 34
When adding, removing, or renaming a skill, update the inventories or structural guidance in AGENTS.md, CLAUDE.md, CODEX.md, and README.md, …
- Distinctive: line 38
When adding a new skill, follow the same pattern used by existing skills:
- Very specific: line 40
Create skills/<skill-name>/SKILL.md with valid YAML frontmatter containing only name and description.
- Very specific: line 41
Add skills/<skill-name>/README.md for user-facing documentation.
- Very specific: line 42
Add practical examples under skills/<skill-name>/examples/, or use docs/ and references/ when the skill is primarily a reference package.
- Very specific: line 43
Add skills/<skill-name>/agents/openai.yaml for optional OpenAI host UI metadata.
- Very specific: line 44
Keep guidance concise, keep SKILL.md under 500 lines, and avoid duplicating large sections between files.
- Very specific: line 50
Check that SKILL.md, its README.md, and supporting examples give consistent guidance.
- Distinctive: line 51
Verify internal links and referenced files exist.
- Distinctive: line 53
Check repository-level skill inventories whenever the set of skills changes.
- Very specific: line 54
Run python3 -m unittest discover -s tests -p 'test_*.py', ruby scripts/validate_skills.rb, and git diff --check before handing off changes.
- Distinctive: line 55
Keep installation code compatible with Python 3.9 or later and test path, conflict, update, and rollback behavior.
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
AGENTS.md
This file contains shared repository guidance for any coding agent or contributor. Host-specific entry points refer here; the skill content does not depend on a particular agent.
Repository overview
This repository is a skills collection focused on Swift and iOS development. Each skill is a self-contained directory with a SKILL.md, a user-facing README.md, and supporting examples or reference material.
Skills in this repo
swiftui-programming-skill- SwiftUI declarative UI developmentswift-modern-architecture-skill- Swift 6 and iOS 18+ architecture patternsios-accessibility-skill- VoiceOver, Dynamic Type, and accessibility best practicesswift-performance-optimization-skill- Performance profiling and optimization patternscross-platform-app-development-skill- Multi-platform Apple app strategiesswift-unit-testing-skill- Swift Testing and XCTest workflowsios-animation-graphics-skill- SwiftUI animation and graphics techniquesmemory-leak-diagnosis-skill- ARC, retain cycles, and memory leak diagnosisswift-SpeechAnalyzer-Framework-Expert- SpeechAnalyzer and SpeechTranscriber for on-device transcription on iOS 26+ and macOS 26+swift-concurrency-migration- Structured Swift 6 concurrency migration and diagnostic workflowsswiftdata-core-data-migrations- SwiftData and Core Data schema, store, and coexistence migrationsapp-intents-widgets- App Intents, App Shortcuts, WidgetKit, and system-surface integrationiphone-duo-design- Adaptive iPhone Duo layouts, controls, continuity, scenes, camera, and readiness verification
Working conventions
- Keep core instructions independent of host tool names, slash commands, plugins, and machine-specific paths. Put host setup in
docs/agent-compatibility.mdor packaging metadata. Describe file reads and validation in terms any capable agent can follow. - Keep skill changes scoped: update only the relevant skill folder unless a cross-cutting fix is required.
- Preserve each skill structure:
SKILL.md,README.md, and supporting content inexamples/,docs/, orreferences/as appropriate. - Prefer modern Swift patterns in examples (Swift Concurrency, Observation, SwiftData when appropriate).
- Verify code snippets remain coherent and compile-ready when copied into a project.
- State minimum Swift, Xcode, and Apple platform versions when guidance depends on newer APIs. Do not apply one repository-wide deployment target to every skill.
- Use lowercase kebab-case for new skill directory names. Do not rename an existing published skill without a compatibility and migration plan.
- When adding, removing, or renaming a skill, update the inventories or structural guidance in
AGENTS.md,CLAUDE.md,CODEX.md, andREADME.md, plus plugin metadata when its description or version is affected.
Adding or updating skills
When adding a new skill, follow the same pattern used by existing skills:
- Create
skills/<skill-name>/SKILL.mdwith valid YAML frontmatter containing onlynameanddescription. Put positive triggers, exclusions, and scope boundaries indescriptionbecause hosts use it for implicit activation. - Add
skills/<skill-name>/README.mdfor user-facing documentation. - Add practical examples under
skills/<skill-name>/examples/, or usedocs/andreferences/when the skill is primarily a reference package. - Add
skills/<skill-name>/agents/openai.yamlfor optional OpenAI host UI metadata; its default prompt should name$<skill-name>. - Keep guidance concise, keep
SKILL.mdunder 500 lines, and avoid duplicating large sections between files.
Validation checklist
- Confirm
SKILL.mdfrontmatter parses and contains exactlynameanddescription. - Confirm descriptions explain both when the skill should and should not activate.
- Check that
SKILL.md, itsREADME.md, and supporting examples give consistent guidance. - Verify internal links and referenced files exist.
- Review Swift snippets against the minimum versions declared by that skill; use a temporary project when compilation matters and no shared build project exists.
- Check repository-level skill inventories whenever the set of skills changes.
- Run
python3 -m unittest discover -s tests -p 'test_*.py',ruby scripts/validate_skills.rb, andgit diff --checkbefore handing off changes. Runbash scripts/typecheck_examples.shwhen Swift examples change, using the declared Apple toolchain. - Keep installation code compatible with Python 3.9 or later and test path, conflict, update, and rollback behavior. Preserve existing published identifiers; normalize only installed copies when a host requires strict names.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

