agent-skills

Reference catalog of SwiftUI, Swift Concurrency (Swift 6 / MainActor-default), SwiftData, WidgetKit, and App Intents antipatterns. Consult BEFORE writing non-trivial SwiftUI code, when a build hits a concurrency or isolation diagnostic, when an animation or transition isn't behaving, when a widget reads stale data, or when reaching for `TimelineView` / `withAnimation` / `matchedGeometryEffect` / custom gesture composition. Use proactively the moment a task touches one of: SwiftUI animation, SF Symbols, navigation, scrolling, SwiftData fetching, App Intents, widget timelines, `@MainActor` boundaries, `Sendable` errors, or `Task { }` lifecycle.

Use when ready to commit and push completed work — when the user says "ship it", "ship this", "commit and push", or otherwise signals work is done and should land. Verifies the change, stages deliberately, drafts a conventional commit, and pushes safely, with explicit guards against shipping broken code, blanket-staging unrelated work, formatter hooks sweeping the tree, fabricated values, and force-resolved pushes.

Use this skill when producing creative, subjective, or aesthetic output where the user likely cannot pre-specify what they want — taglines, names, copy, tone choices, layouts, design directions, naming, structural sketches, anywhere taste is the gating factor. Produce multiple low-fidelity variations early, vary them on dimensions that actually matter, and use the user's reactions to converge — not their pre-specifications. Do NOT trigger when the user has fully specified what they want, when there's clear time pressure for a single answer, or when the brief itself is unclear (clarify first).

Use this skill when the user's request is vague, ambiguous, multi-interpretable, or when the user explicitly invokes a "reference interview" or asks to be interviewed before proceeding. Adapts the librarian's "reference interview" — a structured way of clarifying what someone actually needs before producing an answer. Trigger on open-ended help requests ("help me with X", "I need to write...", "how should I approach..."), topic-only inputs ("Python", "Q3 strategy"), references to undefined context ("the project", "what we discussed"), strategic or high-stakes deliverables, and anywhere the surface question may not match the real underlying need. When unsure whether to trigger, default to triggering. Do NOT trigger only when the user explicitly said "don't ask, just do it", the request is fully specified (audience, purpose, format, and constraints all explicit), or the missing context is purely a taste preference (which hands off to iterative-design).