Use when user references something they recently copied, pasted, or had on their clipboard — explicitly OR implicitly. Trigger on: Implicit references in any language: "the thing I just copied" / "我刚才复制的" / "上次复制的那段" / "earlier I copied" / "the JSON from before" / "that error message" / "那个链接" / "私がコピーしたもの" / "방금 복사한"…
Initialize or update CLAUDE.md coding standards for any project by auto-detecting tech stack and assembling rules from the claude-rules template library. Use when user says 'init rules', 'generate CLAUDE.md', or wants to set up coding standards.
Initialize or update CLAUDE.md coding standards for any project by auto-detecting tech stack and assembling rules from the claude-rules template library. Use when user says "init rules", "初始化规范", "生成 CLAUDE.md", or wants to set up coding standards.
Generate professional app icons for macOS, Windows, iOS, Tauri, and Electron projects. Use when the user asks to: create an app icon, generate app icons, make a logo for their app, design an application icon, convert SVG to app icons, generate .icns/.ico files, or needs icons for Tauri/Electron/desktop apps. Handles…
Adversarial review — think like an attacker or chaotic user to find bugs, edge cases, and failure modes that normal development misses. Use after completing a feature, refactor, or complex bug fix to stress-test the code before it ships. Use when user says: "break-it", "对抗式审查", "搞崩它", "adversarial review", "帮我找漏洞"…
Use when the user wants a single one-command script to start a project's frontend and backend together for local development. Triggers: "生成启动脚本", "帮我生成启动脚本", "写一个dev.sh", "写启动脚本", "前后端一起启动", "一键启动前后端", "一键启动脚本", "generate dev script", "create launch script", "one command to run frontend and backend".
Produce deliverable-grade business flowcharts and technical architecture diagrams as hand-written SVG inside a self-contained HTML page — draw.io visual grammar (standard shapes, line icons, mandatory legend), a preview toolbar (zoom / grid / PNG@2x / clean export mode), and a required render-and-inspect verification…
Upload images to your own GitHub repo and get jsDelivr CDN, GitHub raw, and Markdown URLs back. Supports file paths (single or batch in parallel) and clipboard (screenshots from Cmd+Shift+4, or files copied from Finder). Use when the user says: "upload image", "image upload", "image hosting", "screenshot upload"…
Batch update multi-language localization files in parallel. Use when user says: "更新多语言", "本地化更新", "翻译到所有语言", "update all languages", "localize", "多语言同步", or asks to apply a change across multiple language files.
Browse installed macOS applications and uninstall any of them together with every leftover they left behind — a scriptable replacement for AppCleaner / CleanMyMac. Also use this to simply LIST what is installed, since choosing what to remove usually starts with browsing the full inventory. Scans /Library, /Library…
A macOS maintenance workflow based on tw93/mole. Iron rule: preview first, confirm, then execute. Any deletion requires explicit user text confirmation.
Scan macOS for zombie/stuck processes (orphan MCP servers, leftover Docker cagent, long-running dev servers, old claude sessions, long-lived terminal tabs, large memory over-age) and produce a tiered diagnostic report + suggested kill commands. Trigger words "cleanup zombies", "scan zombies", "system cleanup", "check…
Scaffold a production-ready native macOS app project (SwiftUI + SwiftData + SPM) with auto-update via GitHub Releases, dev/release build scripts, in-app localization, menu-bar persistence, and a landing-page template. Use when the user says: "new macos app", "scaffold mac app", "create macos app", "swiftui scaffold"…
Vet npm / yarn / pnpm packages with socket.dev before they enter the project, and audit an existing project's full dependency tree. Auto-detects the project's package manager (npm / yarn / pnpm) and uses it for the actual install — the safety check is the wrapper, not the installer. Use when the user wants to add a…
First-principles analysis — force the AI to reason from fundamental facts instead of pattern-matching from training data. Use for debugging (find the real root cause, not the surface symptom), architecture decisions (is this the right approach, or just a familiar one?), and solution design (should we really solve it…
Browse and recommend across all locally installed Claude Code skills (personal + plugin + project). Use when the user wants to see what skills are available, has forgotten what they have, or asks which skill best fits a specific scenario. List-mode triggers: "list my skills", "what skills do I have", "show installed…