Writing & Agent Automation Skills

This skill should be used whenever favicons, app icons, or a site's icon branding are involved. Trigger it when the user asks to "generate favicons", "add a favicon", "create a favicon set", "turn my logo into favicons/app icons", "make icons from this logo/SVG/PNG", "set up site.webmanifest / PWA icons / apple-touch-icon", or "check/audit my favicon". Also use it proactively when building, scaffolding, or finishing any website or landing page (a site needs a correct favicon set), and when producing branding assets from a logo. It drives the realfavicon-mcp MCP tools (realfavicongenerate, realfaviconcheck, realfaviconchangelog) to build a complete favicon and app-icon set plus ready-to-paste markup from one master image, audit a live page's favicon setup, and track RealFaviconGenerator changes.

Orchestrates the 4man crew as a Claude Code agent team to build a feature end-to-end or stand up a new codebase: a precise spec, parallel coding and testing in the requestor's own style, and a security/compliance/correctness review ending in a merge verdict. Applies when the user asks to implement, build, add, wire up, or ship a feature; to scaffold, bootstrap, or start a new project or codebase; or to turn a vague product request into working, tested, reviewed code. Does not apply to trivial one-line edits, pure questions, explanations, or read-only exploration, which are handled directly without the crew.

Before writing or editing any code, prose, docs, commit message, or other text, pull the user's recorded conventions from MemPalace — coding-style, writing-voice, and working-prefs — and match them instead of inventing a default; for prose, also run /humanizer:humanizer to strip AI-writing tells. Use at the start of any writing, coding, or editing task — "write", "implement", "refactor", "draft", "document", "README", "commit message", "rename", "reply". The reliable, model-invoked companion to the UserPromptSubmit style-pull hook, for when that nudge isn't enough.

Ground platform-API answers in live, official documentation instead of training data when developing native applications for Apple platforms (macOS, iOS, iPadOS, watchOS, tvOS, visionOS), Windows, or Android. Use whenever the user writes, reviews, debugs, or asks about platform-SDK code, OS-specific APIs, UIKit/SwiftUI/AppKit/WatchKit, WinUI/Win32/.NET/WinRT, Android/Jetpack/Kotlin APIs, App Store/Play Store/Microsoft Store submission questions, minimum-OS-version or deployment-target compatibility, or "which native API should I use for X" on any of these platforms — even when the platform isn't named explicitly but the code clearly targets one (import UIKit, import SwiftUI, using Microsoft.UI, androidx., .xcodeproj, .xcworkspace, AndroidManifest.xml). Trigger even for platform questions that feel "well known" — OS APIs change every release (WWDC, Build, Google I/O) and training data goes stale fast, so check the live docs MCP before answering from memory.