petekp/agent-skills
Overview
This skill helps designers and developers build native-feeling macOS apps with SwiftUI or AppKit, focusing on system conventions, keyboard-first workflows, and macOS visual language like Liquid Glass. It prioritizes making apps behave like a “good Mac citizen”: fast, accessible, multi-window aware, and consistent with platform expectations. Use it to audit or plan menu structure, shortcuts, sidebars, toolbars, icons, and overall app archetype.
How this skill works
The skill inspects your app’s UI and interaction surface against macOS conventions: menu bar layout, keyboard shortcuts, window behavior, sidebar/navigation patterns, toolbar grouping, and where Liquid Glass is applied. It flags deviations from platform standards, recommends keyboard mappings and menu placement, and produces concrete deliverables to resolve gaps (archetype, information architecture, command map, state model, accessibility plan).
When to use it
- Early design phase to choose an app archetype (document, library+editor, utility, menu-bar, pro tool)
- Before implementation to define menu structure and keyboard command map
- While designing sidebars, toolbars, or Liquid Glass usage to avoid misapplication
- When auditing an existing macOS app for platform consistency and accessibility
- Prior to release to confirm multi-window behavior, fullscreen support, and system shortcut compatibility
Best practices
- Prefer built-in system components and conventions; custom chrome must be justified
- Put Settings in the App menu and expose primary commands in the menu bar with standard ⌘ shortcuts
- Make every primary command reachable by keyboard; do not override common system shortcuts (⌘C/⌘V/⌘Z)
- Use Liquid Glass only for navigation/controls (toolbars, sidebars), never for main content layers
- Group toolbar items by function and demote secondary actions into a “more” menu or popover
- Plan accessibility and state early: VoiceOver labels, full keyboard navigation, Reduced Motion, undo/redo, and persistence
Example use cases
- Designing a mail or notes app that uses a sidebar + editor layout and multiple windows
- Converting a single-window utility into a multi-window, document-based workflow
- Auditing an app that misuses translucent chrome or places settings outside the App menu
- Defining a keyboard-first command map for a pro tool with dense shortcuts
- Planning iconography and SF Symbols usage for consistent macOS look-and-feel
FAQ
Use Liquid Glass for navigation and controls (sidebars, toolbars, bars). Avoid applying it to content areas like documents, tables, and lists.
How should I handle keyboard shortcuts?
Expose every primary command via keyboard, follow macOS standard shortcuts, and never reassign common system shortcuts like ⌘C/⌘V/⌘Z.
When should I support multiple windows?
Support multiple windows whenever users work on multiple items or compare content; single-window is appropriate only for simple utilities or menu-bar apps.
30 skills
This skill helps you design macOS apps that feel native and citizen-friendly by aligning with system conventions and Liquid Glass rules.
This skill creates interactive tuning panels to adjust parameters visually, surfacing relevant values and exporting changes for lightweight LLM consumption.
This skill guides you through automated verification of current work, surfacing what Claude can test and prompting you for manual checks when needed.
This skill turns architectural assessments into a safe, verifiable refactor by executing plan-driven chunks with progress tracking and automated verification.
This skill helps you craft precise research prompts for deep research agents by interviewing you to extract context and produce a powerful prompt.
This skill crafts a literate guide that weaves code and prose into a single narrative, clarifying what the code does and why.
This skill helps identify and safely remove dead code, orphaned artifacts, and redundant implementations in agent-maintained codebases.
This skill helps remove AI writing patterns and restore human voice in prose, making AI-assisted text sound natural and engaging.
This skill guides you through a structured deep-work workflow, creating research and planning artifacts before any non-trivial code changes.
This skill ruthlessly analyzes architectural seams to expose coupling, abstraction leaks, and design failures, delivering uncompromising redesign proposals.
This skill conducts exhaustive, citation-rich research across sources, cross-checking facts and presenting structured findings with confidence levels and
This skill helps you reveal hidden system behavior with interactive visualizations of state machines, data flows, and event systems.
This skill exposes structured logs and telemetry to coding agents, enabling runtime visibility and queries for debugging and observability.
This skill generates a structured, evidence-based study guide to onboard newcomers and explain a codebase from purpose to patterns.
This skill creates animated, self-contained explainer visuals as HTML/CSS/JS to clarify ideas in essays and blog posts.
This skill bundles current work into a self-contained review package with a contextual README and a ready-to-use review prompt.
This skill synthesizes outputs from multiple AI models, verifies claims against the codebase, and produces a reliable, evidence-supported assessment.
This skill converts a completed study guide markdown into an interactive, scroll-driven web app with editorial typography, narrated sections, and progressive
This skill acts as a calm, methodical fixer that quickly identifies agent-induced issues and provides a clear path to unblock development.
This skill helps you configure and customize OpenClaw deployments, writing bootstrap files, configuring channels, and building creative automation patterns.
This skill helps you identify and terminate resource-hogging processes to reclaim CPU, memory, and battery while providing a clear hunt report.
This skill pauses work to assess context and surface 2-5 actionable next steps with rationale.
This skill optimizes agent documentation for fast retrieval by building a knowledge manifest, dense compiled artifacts, and task-specific load maps.
This skill analyzes and stabilizes hierarchical matching systems, translating requirements into constraints and guiding architecture, debugging, and
This skill evaluates a codebase for autonomous agent readiness and delivers tailored, actionable recommendations to enable unattended execution.
This skill helps you review proposals methodically by chunking content, predicting feedback, and generating actionable, format-matched insights.
This skill helps you keep inline code documentation clear and co-located with code, explaining why decisions were made and guiding future readers.
This skill applies Model-First Reasoning to code tasks, ensuring a formal model guides implementation for correctness and reduced hallucinations.
This skill analyzes recent conversation to capture concrete learning for project knowledge or cross-project skills and generates actionable notes.
This skill helps you automate Unix and macOS workflows with robust shell scripts and practical system administration guidance.
