team-attention

Run a full implementation verification pass after code or data changes. Use when the user asks to verify, QA, smoke test, run checks, validate a feature, inspect a local app in the browser, capture screenshots, or turn discovered QA issues into regression tests/checklists with user approval.

Hoyeon execution workflow for Codex. Use when the user invokes "$hoyeon-execute" or wants to execute a Hoyeon plan.json through the Bash-first Codex adapter. This adapter loads the canonical execute skill and follows its Codex runtime surface.

Plan-driven orchestrator. Reads plan.json (from /blueprint) or requirements.md, then dispatches workers to build the system. Use when: "/execute", "execute", "plan 실행", "blueprint 실행"

Recon-first browser automation. Orchestrator explores the site first via chromux, saves a guide file with insights, then delegates execution to browser-explorer agent. Use when: "/browser-work", "브라우저 작업", "사이트에서 해줘", "웹에서 해줘", "LinkedIn에서", "크롬으로", "browser task", "automate this site".

Deep web research skill using parallel subagents + chromux browser-explorer + Gemini. Spawns multiple WebSearch research agents AND browser-explorer agents (via chromux for JS-heavy/dynamic sites), plus a Gemini CLI deep research source, then synthesizes everything into a cited report. Uses WebSearch, WebFetch, chromux browser-explorer, and Gemini CLI. Invoke with /deep-research <topic>.

Systematically QA test any application — web apps, native macOS apps, Electron apps, CLI tools, interactive REPLs, or anything on screen. Three modes: browser (chromux/CDP, fast, DOM-level), computer (MCP computer-use, screenshot + pixel clicks, any app), and cli (tmux, send-keys + capture-pane for interactive terminals). Auto-selects mode or accepts --browser / --computer / --cli override. Use when asked to "qa", "QA", "test this site", "test this app", "find bugs", "test and fix", "fix what's broken", "dogfood", "exploratory test", "bug hunt", "QA this app", "사이트 테스트", "앱 테스트", "브라우저 QA", "화면 보고 테스트해줘", "네이티브 앱 테스트", "screen test". Three tiers: Quick (critical/high only), Standard (+ medium), Exhaustive (+ cosmetic). Produces before/after health scores, fix evidence, and a ship-readiness summary.

Hoyeon blueprint workflow for Codex. Use when the user invokes "$hoyeon-blueprint" or wants to turn requirements.md into a validated plan.json and optional contracts.md. This adapter loads the canonical blueprint skill and follows its Codex runtime surface.

Hoyeon browser automation workflow for Codex. Use when the user invokes "$hoyeon-browser-work" or asks Codex to perform a browser task through recon-first chromux automation. This adapter loads the canonical browser-work skill and follows its Codex runtime surface.

This skill should be used when the user asks to "트윗 가져와", "트윗 번역", "X 게시글 읽어줘", "tweet fetch", "트윗 내용", "트윗 원문", or provides an X/Twitter URL (x.com, twitter.com) and wants to read, translate, or analyze the tweet content. Also useful when other skills need to fetch tweet text programmatically.

Generate Korean podcast episodes from any source (URLs, tweets, articles, PDFs) — analyzes content, writes a script, generates audio via OpenAI TTS, converts to MP4, and auto-uploads to YouTube. Use this skill whenever the user says 'make a podcast', 'convert to podcast', 'podcast', 'create an episode', 'turn this into audio', 'YouTube podcast', 'turn this article into a podcast', 'publish as audio', or provides sources and wants them transformed into a listenable format. Supports partial execution: script-only, TTS-only, or upload-only.

Analyze tasks and dynamically assemble expert agent teams using Claude Code's TeamCreate API. Scouts your codebase, designs optimal agents, and executes with validation.

This skill should be used when the user is building, planning, or strategizing and the key question is whether to optimize content (what) or change form (how/medium). Trigger on "내용 vs 형식", "content vs form", "metamedium", "형식을 바꿔볼까", "새로운 포맷", "관점 전환", "perspective shift", "다른 방법 없을까", "같은 방식이 안 먹혀", "diminishing returns". Applies Alan Kay's metamedium concept to surface form-level alternatives. For requirement clarification use vague; for strategy blind spots use unknown.

This skill should be used when the user provides a strategy, plan, or decision document and wants to surface hidden assumptions and blind spots using the Known/Unknown 4-quadrant framework. Trigger on "known unknown", "4분면 분석", "blind spots", "뭘 놓치고 있지", "뭘 모르는지 모르겠어", "전략 점검", "전략 분석", "assumption check", "가정 점검", "quadrant analysis", "what am I missing". Strategy-level blind spot analysis with hypothesis-driven questioning. For requirement clarification use vague; for content-vs-form reframing use metamedium.

This skill should be used when the user's request or requirement is ambiguous and needs iterative questioning to become actionable. Trigger on "clarify requirements", "refine requirements", "요구사항 명확히", "요구사항 정리", "뭘 원하는 건지", "make this clearer", "spec this out", "scope this", "/clarify". Turns vague inputs into concrete specs. For strategy blind spots use unknown; for content-vs-form reframing use metamedium.

This skill should be used when the user asks to "check email", "read emails", "send email", "reply to email", "search inbox", or manages Gmail. Supports multi-account Gmail integration for reading, searching, sending, and label management.

This skill should be used when the user asks to "카톡 보내줘", "카카오톡 메시지", "KakaoTalk message", "채팅 읽어줘", "~에게 메시지 보내줘", or needs to send/read messages via KakaoTalk on macOS.

YouTube 콘텐츠의 챕터별 하위페이지 구조를 readings.ts에 생성하고, 전체 번역 파일에서 챕터별 번역 파일을 자동 생성합니다. 사용: /split-youtube-chapters week1/deep-dive-llms

코드 수정 완료 후 conventional commit 형식의 커밋 메시지를 자동 생성하고 사용자 확인 후 커밋합니다. 사용자가 커밋해줘 또는 /commit 요청 시 사용.

readings.ts의 요약 필드가 원본 kr 마크다운과 비교하여 잘 작성되었는지 평가합니다. 이슈별로 개선사항을 제안하고 사용자 컨펌 후 readings.ts를 수정합니다. 사용: /eval-summary week1 또는 /eval-summary week1/slug

나노바나나 프롬프트의 정확성을 한글 번역본과 비교하여 검증하고, WebSearch로 팩트체크한 후 이슈별로 사용자 확인을 거쳐 수정합니다. 사용: /improve-prompt week1/deep-dive-llms/tokenization

Reading 원문에서 나노바나나 프로용 치트시트 프롬프트를 생성합니다. 3단계 에이전트 파이프라인(분석→구조설계→프롬프트생성)으로 콘텐츠 유형에 맞는 프롬프트 생성. 사용: /nanobanana week1/slug

나노바나나에서 생성된 치트시트 이미지를 public에 복사하고 readings.ts에 cheatsheetImage 필드를 추가합니다. 사용: /publish-cheatsheet week1/slug

나노바나나 치트시트 이미지를 Gemini Vision으로 분석하여 해당 kr md 파일의 내용과 일치하는지 검증합니다. 기술적/문맥적 오류, 잘못된 텍스트/그래프/그림을 찾아냅니다. 사용: /review-cheatsheet week1/slug/chapter

번역된 한국어 콘텐츠의 품질을 AI로 검증합니다. 원문과 번역본을 비교하여 정확성, 완전성, 자연스러움, 용어 일관성을 평가합니다. Claude, Codex, Gemini CLI 3개 AI로 교차 검증하여 합의된 이슈를 우선 표시합니다. 사용: /review-translation week1/slug

Analyze live W&B Models and Weave project data using the W&B MCP server, including runs, histories, configs, traces, evaluations, artifacts, registries, comparisons, diagnoses, and reports. Use when the user asks what happened in a W&B project, which run or model is best, why a run or trace failed, how evaluations performed, what changed between versions, or to save an evidence-backed analysis as a W&B Report.

Add W&B Models experiment tracking and guarded alerts to iterative experiment workflows, including Codex or Ralph-loop autoresearch. Use when the user asks to instrument experiments, create one W&B run per candidate, group related runs, record hypotheses and code provenance, log metrics and summaries, detect failed or stalled experiments, or validate that runs are comparable and reproducible. Use wandb-project-analyst for analysis of data already stored in W&B.

Add W&B Weave tracing to an existing LLM, agent, RAG, or tool-using application with minimal behavioral change. Use when the user asks to integrate Weave, instrument code with @weave.op, trace model or tool calls, choose trace boundaries, account for framework auto-patching, protect sensitive trace data, debug missing traces, or verify a real trace tree. Keep this skill focused on tracing; route evaluations, assets, and general W&B documentation elsewhere.

Turn a world-model, 월드모델, simulation, predictive-state, embodied-agent, or intervention-model idea into ranked falsifiable research questions and a compact experiment handoff. Use when selecting a world-model research direction, deciding whether an idea needs a world model, defining an action-conditioned or counterfactual comparison, or preparing Ralphthon work for Auto Research Track 1 or Track 2.

Use when preparing Ralphthon research specifications or 연구 명세, automated experiments, a Track 1 short paper or 논문, a Track 2 review agent or 리뷰, or an evidence-backed submission under the event deadline.

Use when signing up or 가입 for current VESSL Cloud, installing or authenticating vesslctl, checking organizations and teams, or preparing a first billable GPU Workspace, Job, or 기본 세팅.

Use when signing up or 가입 for W&B Cloud, configuring a personal API key, checking authentication, or creating a first experiment-tracking run or 실험 추적 project.

Welcome and orient Ralphthon @ICML attendees, operators, sponsors, and QR/POP traffic. Use when the user asks for Ralphthon ICML onboarding, attendee welcome copy, event orientation, QR or POP copy, guestbook engagement, hackathon context, or Codex skill surface framing.

길드 운영 인사이트를 구조화된 문서로 축적하여 조직 지식을 복리로 성장시킨다

This skill should be used when the user is building, planning, or strategizing and the key question is whether to optimize content (what) or change form (how/medium). Trigger on "내용 vs 형식", "content vs form", "metamedium", "형식을 바꿔볼까", "새로운 포맷", "관점 전환", "perspective shift", "다른 방법 없을까", "같은 방식이 안 먹혀", "diminishing returns". Applies Alan Kay's metamedium concept to surface form-level alternatives. For requirement clarification use vague; for strategy blind spots use unknown.

This skill should be used when the user provides a strategy, plan, or decision document and wants to surface hidden assumptions and blind spots using the Known/Unknown 4-quadrant framework. Trigger on "known unknown", "4분면 분석", "blind spots", "뭘 놓치고 있지", "뭘 모르는지 모르겠어", "전략 점검", "전략 분석", "assumption check", "가정 점검", "quadrant analysis", "what am I missing". Strategy-level blind spot analysis with hypothesis-driven questioning. For requirement clarification use vague; for content-vs-form reframing use metamedium.

This skill should be used when the user's request or requirement is ambiguous and needs iterative questioning to become actionable. Trigger on "clarify requirements", "refine requirements", "요구사항 명확히", "요구사항 정리", "뭘 원하는 건지", "make this clearer", "spec this out", "scope this", "/clarify". Turns vague inputs into concrete specs. For strategy blind spots use unknown; for content-vs-form reframing use metamedium.

작업을 마무리하는 스킬. 테스트 확인, 커밋/PR, worktree 정리와 함께 5개의 전문 에이전트로 세션 인사이트를 추출합니다.

Create, manage, or talk to an openclone "clone" — a named AI persona with one or more categories (vc, tech, founder, expert, influencer, politician, celebrity) and attached knowledge. Triggers on phrases like "create a clone", "make a persona", "talk as <name>", "switch to <name>", "feed knowledge to", "ingest url for <clone>", "update <name> with latest info", "<name> 클론 최신 정보로 업데이트", "ask all VCs", "stop being <name>", or any mention of `/openclone`. Renders the home panel with no args; name/number activates a clone; sub-actions `stop`, `new`, `ingest`, `update`, `room`, `panel`.

Use when the user wants to consult an AI persona "clone" for advice, strategy, analysis, or domain expertise—especially in startup, VC, tech, growth, HR, or business contexts. Available official clones include: 장동욱(Brian, VC), 김철우(VC/founder), 권도균(Douglas, founder/VC), 조여준(Ethan, VC), 정구봉(tech/founder), 김동현(iid, HR), 신재명(Jay, founder), 조쉬(founder/expert), 이경훈(founder/VC), 김용훈(Levi, growth/CMO). Also use when users ask how to install openclone via npm, run openclone outside Claude Code, choose a provider (OpenAI-compatible API, Codex OAuth, Claude Code subscription OAuth, or local Ollama), pick a model, list or resume saved conversation sessions (openclone history, --resume, --resume=<id>), use --no-persist, understand interactive chat mode, knowledge lookup, conversation persistence under ~/.openclone/conversations, or fix CLI errors.

Use when the user wants to run a long-horizon Codex `/goal` (or any persistence-loop agent) against a Ralph-style repo and needs the goal *shaped* — stop condition pinned, qualitative outcome turned into a verifiable checklist, proxy fidelity declared, and durable memory files seeded. Trigger on "shape my goal", "set up /goal", "make this a /goal", "/goal from RALPH.md", "long-running goal", "ralph goal", "goal prompt", "stop condition", or whenever a user asks to convert a vague long-running task into a Codex `/goal` prompt. Reads `RALPH.md` when present, fills the four shape checks (Outcome / Stop / Proxy / Memory), refuses to emit a goal that would loop forever, and writes seed `PLAN.md` / `EXPERIMENTS.md` / `EXPERIMENT_NOTES.md` next to `RALPH.md` so the loop has a place to remember.

Use in Codex Plan mode when the user wants to clarify, shape, explore, visualize, or expand an idea through request_user_input choice questions and AI-generated image previews. Trigger on "clarify-image", "visualize my idea", "turn my idea into an image", "help me think through this visually", visual brainstorming, image-first ideation, or when the user wants Codex to ask structured multiple-choice questions, turn the answers into image previews, and use those images to make the idea clearer and stronger. This is a Plan-mode, request_user_input-based skill for Socratic idea clarification plus image generation, not for software implementation requirements.

SDD workflow executor - orchestrator-controlled spec-driven development

SDD workflow executor - orchestrator-controlled spec-driven development. Activate when the user wants to run spets, SDD workflows, or spec-driven development tasks.

SDD workflow executor - orchestrator-controlled spec-driven development. Activate when the user wants to run spets, SDD workflows, or spec-driven development tasks.

Interactive markdown review with web UI. Use when: - User says "review this", "check this plan", "/review" - User wants to give feedback on a plan or document - User says "let me review", "피드백", "검토해줘" - User specifies a file path to review (e.g., "review /path/to/file.md")