setup

Run forever as a resident worker in ONE Codex turn: every 5 minutes poll a watched directory for new task files and execute each one to completion, then keep watching — the turn never ends. Use when the user says "queue-loop", "무한 루프로 돌려", "큐 감시해서 계속 실행", "resident worker", "watch a directory and run new tasks forever", or wants a single invocation that keeps picking up dropped commands indefinitely. Because the work stays inside one already-running turn, it keeps issuing model requests across usage-limit windows (the workspace spend cap gates only NEW top-level turns, not an in-flight one). Do NOT use for a single one-shot task, or when the user wants the turn to finish and report back.

Use when the user wants Korean prose cleaned of AI tone, stiff old-fashioned (할아버지/훈장/공문서) tone, translationese, and excessive quotation marks - or asks to "글 다듬어", "자연스럽게", "AI 말투 제거", "할아버지 말투 빼", "deslop 한국어", or "writing-fixer". Removes as many AI and unnatural expressions as possible in one thorough pass while preserving meaning, code, math, and markup. Korean prose only.

Inject a task into the queue-loop worker's queue so a running Codex resident worker picks it up on its next poll (~5 min). Use when the user says "queue-submit", "큐에 작업 넣어", "task 주입해", "이거 워커한테 넘겨", "submit a task to the codex worker", "enqueue this for queue-loop", or wants to hand a job off to the always-on worker instead of doing it inline. It only writes a *.task file into the watched queue directory (default ~/codex-queue) — it does NOT run the task in this session. Do not use when the user wants the work done right now in the current session.

한국어 텍스트에서 AI가 쓴 흔적, 번역체, 부자연스러운 표현을 제거하고 자연스러운 한국어로 다시 써 주는 스킬. 말투 통일, 문장부호 교정, 피동 표현 정리, 영어 직역체 제거를 수행한다.

Iteratively review the current repository, plan fixes from the reviews, and implement the plans. Runs three prompts in strict order (deep review → plan from reviews → implement with ralph) and repeats the full three-prompt cycle up to N times (default 100). Context is cleaned between iterations by spawning each full cycle in a fresh subagent.

Browser configuration: viewport sizing, device emulation, geolocation, offline mode, color scheme, and HTTP settings.

Browser debugging: execute JavaScript, view console output and page errors, record traces, and run performance profiling.

Browser interaction commands: clicking, typing, filling forms, selecting options, scrolling, dragging, file uploads, and raw mouse/keyboard input.

Network interception, request monitoring, cookie management, localStorage and sessionStorage operations, HTTP headers, and credentials.

Query page content: get text, HTML, attributes, element state, accessibility snapshots, and find elements by role, text, label, placeholder, or testid.

Headless browser automation CLI for AI agents. Core navigation, session management, tabs, frames, and dialogs using the agent-browser CLI.

Browser authentication state management: save and restore login sessions, cookies, and storage across automation runs.

Visual capture and comparison: screenshots, PDFs, element highlighting, accessibility snapshot diffs, screenshot diffs, and URL comparisons.

Wait for elements, timeouts, text appearance, URL patterns, page load states, and custom JavaScript conditions.

Fetch web content, download files, and make HTTP requests. Replaces the mcp-server-fetch MCP server using built-in WebFetch and curl via Bash.

File and directory operations using built-in tools. Replaces the @modelcontextprotocol/server-filesystem MCP server. Maps all 14 filesystem MCP tools to their native equivalents.

Git version control operations using the git CLI. Replaces the mcp-server-git MCP server with direct git commands via Bash.

GitHub platform operations using the gh CLI. Replaces the GitHub MCP server (Docker-based ghcr.io/github/github-mcp-server) with native gh commands.

Browser automation and testing using Playwright CLI and scripts. Replaces the @playwright/mcp MCP server with direct npx playwright commands and temporary script files.