Agent Skills

Reusable agent skills for Apple platform engineering, product work, and project workflows. The active catalog contains 112 skills, grouped into plugin bundles but usable as plain SKILL.md directories by any agent runtime that supports local skills.

What This Repo Contains

• Active skills live in plugins/<bundle>/skills/<skill-name>/SKILL.md.
• Long guidance, examples, and checklists live under each skill's references/ directory.
• Project workflow personas live in plugins/project-tools/agents/.
• Plugin metadata lives in .claude-plugin/ directories for runtimes that install bundled plugins.
• Repository validation helpers live in scripts/.
• example-setup/ contains compact, generic project instruction and documentation templates.

Lifecycle View

The lifecycle is exposed as skills, not command wrappers. Use next-step-router when the next phase is unclear.

| Phase | Primary Skills | Use | |---|---|---| | Define | interview-me, idea-refine, spec-workflow, design, user-test | Clarify intent, explore ideas, scope product direction, and define acceptance criteria before building. | | Plan | plan, project-state-audit, blast-radius, source-driven-development | Turn scope into ordered work, find affected systems, and verify source-backed API assumptions. | | Build | implement, incremental-implementation, test-driven-implementation, small-change | Execute approved work with the right level of slicing, tests, and verification. | | Doubt | doubt-driven-development, strategy-audit, independent-review | Stress-test non-trivial implementation hypotheses, strategy claims, and high-risk decisions. | | Debug | hypothesis-debug, deep-investigate, forensics, agent-harness-audit | Reproduce failures, isolate causes, fix bugs, or improve missing agent harnesses. | | Review | code-review, code-simplification, code-style, ux-audit, pre-ship | Review correctness, quality, complexity, user experience, and release readiness. | | Ship | post-implementation-qa, pr-shipping, change-summary, session-handoff | Run validation, summarize changes, commit, push, open PRs, and leave durable handoff state. |

Catalog Summary

| Bundle | Skills | Purpose | |---|---:|---| | plugins/ios-frameworks | 35 | Apple framework skills for iOS and Apple platform development. | | plugins/ios-platform | 11 | Cross-cutting Apple platform skills for architecture, debugging, testing, tooling, HIG, Swift language, and App Store release work. | | plugins/swiftui | 9 | SwiftUI skills for implementation, design exploration, hardening, performance, visual review, shaders, and Figma translation. | | plugins/uikit | 6 | UIKit skills for Auto Layout, collection views, Core Animation, navigation, fundamentals, and SwiftUI interop. | | plugins/project-tools | 47 | Generic project workflow skills for planning, specs, debugging, review, refactoring, release, memory, and session hygiene. | | plugins/product | 4 | Product skills for analytics interpretation, App Store marketing, monetization strategy, and onboarding. | | Total | 112 | |

Installation

Codex

Use the skill directories directly, or copy/symlink them into the Codex skills directory.

mkdir -p "${CODEX_HOME:-$HOME/.codex}/skills"
ln -sfn "$PWD"/plugins/*/skills/* "${CODEX_HOME:-$HOME/.codex}/skills/"

Use copies instead of symlinks when you want a pinned snapshot that does not change as this repository changes.

Claude Code

Install the plugin bundles from the Claude plugin marketplace:

/plugin marketplace add 4eleven7/Claude-Skills
/plugin install ios-frameworks@claude-skills
/plugin install ios-platform@claude-skills
/plugin install swiftui@claude-skills
/plugin install uikit@claude-skills
/plugin install project-tools@claude-skills
/plugin install product@claude-skills

Update installed plugins:

/plugin marketplace update claude-skills

Skill Catalog

ios-frameworks (35 skills)

Apple framework skills for iOS and Apple platform development.

| Skill | Use | |---|---| | accessibility-audit | Use when auditing Apple platform UI for VoiceOver, Dynamic Type, keyboard navigation, contrast, semantics, and patch-ready accessibility fixes. | | activitykit | Use when implementing or reviewing Live Activities, Dynamic Island, Lock Screen presentations, or ActivityKit updates. | | alarmkit | Use when implementing or reviewing AlarmKit alarms, countdowns, presentations, authorization, or timer flows. | | app-intents | Use when implementing or reviewing App Intents, App Shortcuts, Siri, AppEntity, EntityQuery, Spotlight, or Control Center actions. | | apple-haptics | Use when designing, implementing, or reviewing haptics with sensoryFeedback, UIFeedbackGenerator, Core Haptics, or AHAP patterns. | | authentication-services | Use when implementing or reviewing Sign in with Apple, passkeys, credential providers, AutoFill, or AuthenticationServices flows. | | background-tasks | Use when scheduling, reviewing, or debugging BGTaskScheduler, background refresh, processing tasks, or SwiftUI background tasks. | | camera-media | Use when implementing or debugging camera capture, AVCaptureSession, PhotosPicker, PHPicker, limited library access, or HDR media handling. | | cloudkit | Use when implementing, reviewing, or debugging CloudKit sync, CKSyncEngine, SwiftData with CloudKit, iCloud Drive, or iCloud key-value storage. | | combine | Use when writing, reviewing, or migrating Combine publishers, subscribers, subjects, operator chains, or async bridges. | | core-data | Use when writing or reviewing Core Data stacks, models, migrations, concurrency, relationships, or CloudKit integration. | | core-location | Use when implementing or debugging Core Location authorization, CLServiceSession, CLMonitor, background location, or async location updates. | | coreml-vision | Use when implementing or reviewing on-device machine learning with CoreML model loading, predictions, Vision image analysis, feature providers, or ML pipeline integration. | | discoverability | Use when planning or debugging app discoverability through App Intents, Spotlight, Siri suggestions, Action Button, or Control Center. | | foundation-models | Use when implementing Apple Foundation Models sessions, guided generation, tool calling, streaming, or availability checks. | | grdb | Use when writing or reviewing GRDB records, SQL, migrations, associations, observation, concurrency, or persistence safety. | | healthkit | Use when implementing or reviewing HealthKit authorization, samples, queries, workouts, background delivery, or HealthKit UI. | | ios-energy | Use when diagnosing battery drain, thermal issues, power impact, polling, timers, Low Power Mode behavior, hot devices, Power Profiler findings, or energy efficiency regressions. | | ios-file-storage | Use when choosing iOS file locations, backup behavior, file protection, disk pressure handling, atomic writes, or file I/O safety. | | ios-localization | Use when implementing or reviewing string catalogs, plurals, RTL, locale-aware formatting, XLIFF, or localized App Shortcuts. | | ios-networking | Use when implementing or reviewing URLSession, Network.framework, async network connections, diagnostics, or iOS networking failures. | | mapkit | Use when implementing or debugging MapKit maps, annotations, clustering, search, directions, Look Around, or camera behavior. | | metal | Use when implementing or debugging Metal render/compute pipelines, shaders, MTKView, command buffers, textures, or GPU capture. | | pdfkit | Use when displaying, loading, annotating, searching, selecting, or bridging PDFKit documents into SwiftUI. | | push-notifications | Use when implementing or debugging APNs registration, permission flows, payloads, notification extensions, Live Activity pushes, or FCM. | | realitykit | Use when implementing or debugging RealityKit, visionOS spatial computing, RealityView, entities, components, systems, AR, or 3D content. | | scenekit | Use when implementing or debugging SceneKit scenes, nodes, materials, lighting, cameras, physics, animation, or SceneView bridging. | | sf-symbols | Use when working with SF Symbols rendering modes, symbol effects, animation, palettes, variable symbols, or accessibility labels. | | spritekit | Use when implementing or debugging SpriteKit scenes, nodes, actions, physics, particles, tile maps, SpriteView, or texture atlases. | | storekit | Use when implementing or reviewing StoreKit 2 products, purchases, transaction verification, subscriptions, entitlements, or tests. | | swift-charts | Use when building or reviewing Swift Charts marks, axes, scales, legends, selection, scrolling, vectorized plots, or chart accessibility. | | swiftdata | Use when writing or reviewing SwiftData models, queries, migrations, schema policy, runtime patterns, or persisted schema changes. | | tipkit | Reviews and writes TipKit code — tip definitions, eligibility rules, events, display frequency, popover and inline tips, and testing. Use when implementing feature discovery tips. | | translation-framework | Use when adding Apple Translation framework UI, custom sessions, language availability checks, or batch translation. | | widgetkit | Use when implementing or debugging WidgetKit timelines, widgets, Live Activities, Control Center controls, or extension lifecycle limits. |

ios-platform (11 skills)

Cross-cutting Apple platform skills for architecture, debugging, testing, tooling, HIG, Swift language, and App Store release work.

| Skill | Use | |---|---| | app-store-release | Use when preparing App Store submission, privacy manifests, metadata, rejection responses, or App Store Connect release work. | | hang-diagnostics | Use when diagnosing app freezes, watchdog terminations, main-thread blocking, MetricKit hangs, or Swift concurrency deadlocks. | | ios-architecture | Use when designing or reviewing iOS app architecture, service boundaries, dependency injection, composition roots, or module guardrails. | | ios-debugging | Use when debugging iOS runtime issues with LLDB, diagnostics, memory tools, stack traces, breakpoints, or SwiftUI state inspection. | | ios-tooling | Use when building, running, testing, profiling, debugging, documenting, or releasing iOS/macOS apps with command-line tooling. | | ios-ux-writing | Use when writing or reviewing iOS microcopy, alerts, empty states, error messages, button labels, or accessibility text. | | platform-hig | Use when building, reviewing, or refactoring UI against Apple Human Interface Guidelines across iOS, iPadOS, macOS, tvOS, watchOS, or visionOS. | | swift-concurrency | Use when writing or reviewing async/await, actors, Sendable, task cancellation, isolation, or Swift concurrency migrations. | | swift-language | Use when writing or reviewing Swift language features, ownership, noncopyable types, InlineArray, Span, typed throws, or C bounds safety. | | swift-testing | Use when writing, reviewing, debugging, or improving Swift Testing code with modern APIs, TDD workflow, fixtures, traits, expectations, or test organization. | | xcuitest | Use when writing or debugging XCUITest flows, accessibility identifiers, waits, page objects, system alerts, or UI test failures. |

swiftui (9 skills)

SwiftUI skills for implementation, design exploration, hardening, performance, visual review, shaders, and Figma translation.

| Skill | Use | |---|---| | figma-to-swiftui | Use when translating Figma designs, nodes, tokens, or assets into production SwiftUI for an iOS app. | | screen-design-exploration | Use when exploring SwiftUI screen design directions, comparing layout variants, or turning an ambiguous screen idea into a concrete visual direction. | | swiftui-harden | Use when hardening SwiftUI views against edge cases: long text, empty data, large datasets, errors, Dynamic Type extremes, RTL, concurrency, accessibility, or pre-ship resilience. | | swiftui-microinteractions | Use when designing, reviewing, or implementing iOS microinteractions, toggles, confirmations, pull-to-refresh, input validation, interactive polish, or signature product moments. | | swiftui-patterns | Use when building, reviewing, or refactoring SwiftUI view code, data flow, composition, navigation, accessibility, performance, polish, or SDK migrations. | | swiftui-performance-audit | Use when diagnosing SwiftUI jank, slow rendering, high CPU/memory, identity churn, layout thrash, or update storms. | | swiftui-shaders | Use when building or reviewing SwiftUI Shader API effects, Metal shader functions, ripple, pixel, color, distortion, or layer effects. | | swiftui-view-refactor | Use when refactoring SwiftUI views, splitting bodies, extracting subviews, removing side effects, or simplifying data flow. | | swiftui-visual-review | Use when reviewing SwiftUI views for visual quality, accessibility, HIG fit, state coverage, polish, and design-system drift. |

uikit (6 skills)

UIKit skills for Auto Layout, collection views, Core Animation, navigation, fundamentals, and SwiftUI interop.

| Skill | Use | |---|---| | auto-layout | Use when building or reviewing UIKit Auto Layout constraints, UIStackView, priorities, hugging, compression, or safe areas. | | core-animation | Use when writing or reviewing CALayer, CAAnimation, UIViewPropertyAnimator, or UIKit animation bridges. | | uikit-collections | Use when building or reviewing UITableView, UICollectionView, diffable data sources, compositional layout, or cell registration. | | uikit-fundamentals | Use when reviewing UIKit lifecycle, view hierarchy, responder chain, traits, size classes, or modernization away from shared state. | | uikit-interop | Use when bridging SwiftUI and UIKit with representables, UIHostingController, UIHostingConfiguration, or incremental adoption. | | uikit-navigation | Use when building or reviewing UIKit navigation controllers, modal presentations, custom transitions, or coordinator routing. |

project-tools (47 skills)

Generic project workflow skills for planning, specs, debugging, review, refactoring, release, memory, and session hygiene.

| Skill | Use | |---|---| | agent-harness-audit | Use when an agent repeatedly fails, needs manual context, cannot reproduce or verify a bug, cannot inspect runtime behaviour, or relies on copy-pasted human knowledge. | | blast-radius | Use when analyzing a diff, branch, or planned change to identify affected files, dependents, tests, release risk, rollback scope, or hidden blast radius before shipping. | | breaking-change-guard | Use when changing behavior, persisted state, schemas, contracts, flags, routing, or architecture where compatibility, migration, fallback, or dual behavior is being considered. | | capture-idea | Use when capturing an early idea or out-of-scope improvement in a lightweight backlog without derailing current work. | | capture-lesson | Use when a correction, gotcha, near-miss, surprising behavior, or repeated mistake should be captured as durable project knowledge. | | change-summary | Use when summarizing changes, documenting root cause, listing verification, producing rollback notes, or drafting a change-summary. | | code-review | Use when reviewing code changes for defects, regressions, missing tests, risky design decisions, security issues, performance problems, or overall quality before shipping. | | code-simplification | Use when working code is harder than necessary to understand, review, maintain, test, or change, and behavior must be preserved while reducing complexity. | | code-style | Use when auditing or fixing non-mechanical code style issues that formatters and linters usually miss, including clarity, naming, consistency, and house-style drift. | | competitor-research | Use when researching competitor apps, product alternatives, market references, app listings, pricing, positioning, or structured competitor entries for a project reference document. | | debate | Use when comparing approaches or tradeoffs through structured opposing arguments before choosing a technical or product direction. | | deep-investigate | Use when diagnosing a specific root cause without applying fixes, performing read-only investigation, building a causal chain, or producing an evidence-backed report. | | design | Use when creative implementation work needs clarified intent, approach exploration, design direction, user-facing product decisions, or approval before code changes. | | doubt-driven-development | Use when a non-trivial implementation decision, hypothesis, safety claim, migration, or architecture choice needs adversarial review before it becomes code or shipping guidance. | | evaluate-findings | Use when validating external feedback, review comments, AI findings, dead-code reports, or suspected false positives. | | explain | Explain how code works with summaries, diagrams, analogies, and annotated snippets. Use when the user asks "how does X work?" or wants a codebase walkthrough. | | find-dead-code | Use when finding unused production code, dead exports, stale helpers, test-only references, or deletion candidates with analysis-only reporting before any removal. | | forensics | Use when investigating what happened after confusing failures, regressions, bad merges, unexpected repository state, unclear causality, or contradictory evidence. | | git-worktrees | Use when starting isolated feature work in a git worktree or when implementation needs branch/worktree separation. | | hypothesis-debug | Use when debugging and fixing a bug end-to-end with reproduction, hypotheses, evidence, a minimal fix, regression coverage, and verification. | | idea-refine | Use when a raw idea, feature concept, product direction, or speculative improvement needs divergent exploration, assumption testing, convergence, and a concrete next-step artifact. | | implement | Use when implementing approved specifications or scoped feature work through a test-first, verified project workflow after requirements and boundaries are clear. | | incremental-implementation | Use when implementing multi-file changes, large tasks, risky edits, or approved plans that need careful slicing, verification after each slice, and rollback-friendly progress. | | independent-review | Use when asking an independent reviewer or external review tool to audit recent changes, then validate each finding before acting. | | interview-me | Use when an ask, idea, feature, plan, or product direction is underspecified and the user wants guided questions, multiple-choice tradeoffs, recommended defaults, or collaborative scope shaping before specs or implementation. | | multi-agent-implementation | Use when executing an implementation plan with multiple mostly independent tasks that can be delegated and reviewed in parallel. | | next-step-router | Use when deciding which workflow or skill should run next based on conversation state, repo state, and verification status. | | plan | Use when work requires a persistent multi-step plan, findings log, or progress file before implementation. | | post-implementation-qa | Use when implementation is complete and the work needs project-defined validation, diff review, lesson capture, or handoff to commit and PR shipping. | | pr-shipping | Use when committing, staging, pushing, opening pull requests, monitoring checks, preparing merge notes, or landing changes through a disciplined PR workflow. | | pre-ship | Use when running a final quality gate across spec compliance, UX, visual polish, verification, release readiness, and unresolved risk before shipping. | | project-skill-audit | Use when auditing, deduplicating, renaming, removing, or improving project skill catalogs. | | project-state-audit | Use when reconciling repository code, specs, scope, todos, and backlog items to determine current project state and next work. | | reference-library | Use when saving useful external references such as repositories, articles, documentation, examples, or research notes into a project reference library. | | session-distill | Use when distilling a session, saving learnings, capturing durable corrections, extracting lessons, updating project memory, or improving future agent behaviour. | | session-handoff | Use when pausing, resuming, or creating a durable handoff so later work can continue safely. | | skill-research | Use when researching a technology or external skill to create, update, or improve reusable agent skills from verified sources. | | small-change | Use when making a trivial or small targeted change that needs direct implementation plus the minimum verification appropriate to its risk. | | source-driven-development | Use when framework, SDK, API, dependency, or platform-specific code must be grounded in current official documentation instead of memory or stale examples. | | spec-workflow | Use when drafting, clarifying, researching, or syncing a feature specification before or after implementation. | | strategic-refactor | Use when finding high-leverage refactors, identifying repo-wide complexity, choosing the best refactor, reducing architectural drag, or ranking refactor candidates. | | strategy-audit | Use when proposing, reviewing, or choosing a strategy, plan, architecture, migration, rollout, or confidence claim. | | test-driven-implementation | Use when implementing specs with significant business logic, persistence, complex rules, or high-risk behaviour where independent behavioural tests should drive the implementation. | | tidy-memory | Use when auditing, shrinking, organizing, compacting, trimming, or deduplicating agent memory or project memory that has become stale or too large. | | user-test | Use when evaluating a feature, idea, spec, design, onboarding flow, product decision, or tradeoff through target personas before implementation or release. | | ux-audit | Use when auditing user-facing workflows, screens, navigation, onboarding, conversion paths, or product promises from entry point through outcome. | | writing-skills | Use when creating new skills, editing existing skills, or verifying skills work before deployment |

Project Personas

Reusable persona prompts live in plugins/project-tools/agents/.

| Persona | Use | |---|---| | code-reviewer | Fresh code review across correctness, architecture, tests, security, performance, and maintainability. | | test-engineer | Test strategy review, regression coverage, edge cases, and verification gaps. | | security-auditor | Practical security review for exploitable risks, trust boundaries, secrets, auth, and privacy. | | web-performance-auditor | Web performance review with metric honesty for Core Web Vitals, loading, rendering, and JavaScript cost. |

product (4 skills)

Product skills for analytics interpretation, App Store marketing, monetization strategy, and onboarding.

| Skill | Use | |---|---| | analytics-interpretation | Use when interpreting product analytics, retention, funnel, cohort, LTV, ARPU, payback, or App Store metric changes. | | app-store-marketing | Use when working on ASO, App Store metadata, keywords, screenshots, previews, customer review replies, Apple Search Ads, or featuring. | | monetization-strategy | Use when choosing monetization, pricing, subscriptions, trials, freemium tiers, paywall splits, ARPU, or charge-readiness. | | product-onboarding | Use when designing or reviewing onboarding, activation, first-run flows, feature adoption, setup tasks, or time-to-value. |

Skill Authoring Rules

• Name skills after what they do, not what they aspire to do.
• Keep frontmatter standard, succinct, and specific: only name and description.
• Keep SKILL.md focused on trigger boundary, routing, workflow, and hard rules.
• Move detailed patterns, examples, and checklists into references/.
• Prefer generic, runtime-neutral wording inside skills.
• Keep framework-specific skills separate when merging would reduce correctness.
• Add routing notes between competing or complementary skills.
• If a workflow needs project setup, search for it first; ask only when it cannot be found.

Validation

Run the local validator before shipping catalog changes:

node scripts/validate-skills.js

The validator checks plugins//skills//SKILL.md frontmatter, description length, trigger wording, required Markdown structure, and dead cross-skill references.

License

MIT