메뉴
Visual capture and comparison: screenshots, PDFs, element highlighting, accessibility snapshot diffs, screenshot diffs, and URL comparisons.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
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.
| name | agent-browser-visual |
| version | 1.0.0 |
| description | Visual capture and comparison: screenshots, PDFs, element highlighting, accessibility snapshot diffs, screenshot diffs, and URL comparisons. |
| allowed-tools | ["Bash","Read"] |
Take screenshots, generate PDFs, highlight elements, and compare visual states.
agent-browser screenshot [path] # Screenshot visible viewport
agent-browser screenshot [path] --full # Full page screenshot
agent-browser screenshot [path] --annotate # Screenshot with element annotations
Default path: ./screenshot.png if omitted.
agent-browser pdf <path> # Save page as PDF
agent-browser highlight <sel> # Visually highlight element on page
agent-browser diff snapshot # Compare current vs last snapshot
agent-browser diff snapshot --baseline <file> # Compare current vs saved baseline
agent-browser diff screenshot --baseline <file> # Visual diff vs baseline image
agent-browser diff url <url1> <url2> # Compare two URLs side by side
# Take a full-page screenshot
agent-browser screenshot /tmp/page.png --full
# Save page as PDF
agent-browser pdf /tmp/report.pdf
# Highlight an element for debugging
agent-browser highlight "#error-message"
# Save baseline snapshot, make changes, then diff
agent-browser snapshot > /tmp/baseline.txt
# ... perform actions ...
agent-browser diff snapshot --baseline /tmp/baseline.txt
# Compare staging vs production
agent-browser diff url "https://staging.example.com" "https://example.com"
--full for pages with scrollable content--annotate to add visual labels to interactive elements/tmp/ for ephemeral captures