shabaraba

Orchestrates the async development workflow (design → develop → review → QA). Handles phase transitions when triggered by workflow-continue hook.

Autonomous development workflow that proactively asks questions, verifies implementation with tools, and auto-fixes linter errors. Use when implementing features to ensure quality through self-driven verification.

Backend development skill for Cloudflare Workers, D1 database, and REST API. Use when implementing API features.

Frontend development skill for React, Next.js, Tailwind CSS, and TypeScript. Use when implementing web frontend features.

iOS development skill for Swift, SwiftUI, Live Activities, WidgetKit, and XCTest. Use when implementing iOS features.

Core orchestration skill for the Claude Organization. Manages async sub-agents, worktrees, daily logs, and context handoff. This skill defines how the main Claude Code instance operates as Secretary to the CEO.

This skill should be used when the user asks about "refactoring", "code smells", "extract method", "extract class", "rename refactoring", "move method", or when improving code structure without changing behavior. Provides Martin Fowler's refactoring catalog and code smell detection.

This skill should be used when the user asks about "code complexity", "cyclomatic complexity", "cognitive complexity", "code metrics", "maintainability index", "code coverage", or when measuring code quality quantitatively. Provides metrics thresholds and measurement techniques.

Create, list, attach to, and finish git-worktree-backed isolated work areas for vibing.nvim chats, entirely via natural language — no separate UI. Use when the user wants to isolate work in its own worktree ("split this into its own worktree", "start this in isolation"), wants to see what worktrees/branches exist ("what worktrees do I have", "what's in progress"), wants to switch/attach the current or a new chat to an existing worktree ("let's go into the auth-fix worktree", "attach to worktree X"), or wants to clean one up when done ("clean up this worktree", "I'm done with this branch's worktree").

Use when working on a project that has a running Neovim instance via vibing.nvim (the vibing-nvim MCP server is connected). Reads live buffer, window, cursor, and selection state through vibing-nvim MCP tools before editing or answering, instead of relying on stale file reads or guesses about what the user currently has open or selected.

Use when navigating or analyzing a codebase that's open in a running Neovim instance via vibing.nvim with an LSP client attached. Prefers the vibing-nvim MCP LSP tools (definitions, references, hover, diagnostics, symbols, call hierarchy) over text search for "where is X defined/used/called", "what type is this", "what's broken here" style questions, since they reflect language-aware, live analysis rather than a text match.

Use when a vibing.nvim chat session's context feels lost or discontinuous — after a session reset, a dropped connection, or when Claude's own reasoning no longer matches what was discussed earlier in this conversation. Re-reads this conversation's own chat buffer (the file whose path is announced via the "Current vibing.nvim chat buffer file:" line in the system prompt) to silently restore context. Also invoke when the user explicitly asks to "remember", "recall", or "re-read the chat history" (in any language). Not for browsing other, unrelated past chat files — use vibing-chat-search for that.

Use when the user wants to find a past vibing.nvim conversation by topic — phrases like "前に〜について聞いたことあったっけ", "did we talk about X before", "find that chat where I asked about Y". Searches every chat file under .vibing/chat/ (both User and Assistant content) for a natural-language query, narrows candidates with grep, then reads and semantically judges the survivors before presenting matches. Not for recovering this conversation's own lost context — use vibing-chat-recall for that.

Automatically design comprehensive E2E test cases for newly implemented vibing.nvim features. Use immediately after completing feature implementation (Phase 5.4) and before running E2E tests. Generates test scenarios covering Happy paths, Error cases, Edge cases, and Integration points with priority ranking (Critical/High/Medium/Low) and ready-to-use test code templates using e2e_helper.lua.

Complete GitHub workflow for Claude Code on the web. ALL GitHub operations MUST use REST API (never gh CLI). Includes branch naming (claude/*-sessionId), push retry logic, PR/issue management via API, and complete workflows. Use for all GitHub interactions in Claude Code web environment.

This skill should be used when starting a large or multi-step task — a multi-file change or migration, a cross-module feature, a production incident, deep research or technology selection — or when work will span many tool calls, sessions, or packages, before code is written or an approach is committed to.

This skill should be used when producing visual or interactive creative work driven by code or configuration — Blender/bpy scenes, game features and game feel, ComfyUI or other image-generation workflows, shaders, procedural art — both before the first version is generated and when critiquing, debugging, or iterating on existing renders, feel, or workflows.