2lab-ai

Convert markdown to a beautifully designed editorial-style PDF report. Uses Pretendard variable font, Swiss-Coffee + Cool-Blue palette, minimalist-maximalism layout, cover page, table of contents, section dividers, page numbers. Optimized for Korean+English mixed reports. Use when user asks for "report PDF", "보고서 PDF", "아름다운 PDF", "예쁜 PDF", or wants to publish a markdown document as a polished document.

블로그 URL 1개를 받아 60초 9:16 숏폼 영상(mp4)을 자동 생성하는 파이프라인. 한국 블로그 4종(티스토리/네이버/벨로그/브런치)에서 본문·이미지를 추출하고, 2인 대화 대본을 만든 뒤 로컬 fish-tts(soma-voice HTTP)와 GPT Image 2(gpt-image-2)로 음성·이미지를 생성해 Remotion v4로 합성한다. "블로그 URL로 숏폼 만들어줘", "이 글로 릴스 영상 만들어줘", "blog to shortform", "blog-shortform", "blog-url-to-shortform" 같은 요청에서 발동한다.

Speech-to-Text using Cohere Transcribe (2B, Apache 2.0)

CLI UX/spec: args, flags, help, output, errors, config, dry-run.

Use when 지혁 asks to record/save/archive a conversation between him and Elon — produces (1) verbatim RAW markdown, (2) cleaned-up EDITED markdown with nothing removed, and (3) a messenger-style HTML blog post. Triggers on "대화 기록해줘", "이거 저장해서 보내줘", "블로그에 올리게 html로", "원문 그대로 + 교정본 + 포스트".

Production frontend UI: pages, apps, components, polished non-generic design.

GitHub deep review: bugs, PRs, best fix, stale-or-real, read-code-first.

100% hand-drawn image generation with Korean Hangul support. Zero font files — all characters rendered from manually-defined stroke primitives. Produces notebook-style illustrated cards with wobble effects. Actions: 손그림, 손글씨 이미지, hand-draw, 일러스트 만들어, 운동 이미지

Use when the user's request is ambiguous — vague asks, polysemous instructions, unclear scope — BEFORE committing to a feature. Two-track loop (one question at a time + parallel codebase exploration) producing a Context Brief that becomes the SSOT for the follow-up work.

Use when code behaves differently from expectations — 'why does this happen', 'find the bug', 'follow the callstack', symptom reports even without the word debugging. Blackbox methodology: AS-IS/TO-BE confirmation, callstack tracing recorded in a CWD-relative trace file, red-green fix.

Use when a trace.md with unfinished scenarios exists and the user wants autonomous implementation — 'continue the work', 'implement the traced feature', or after a bundle was selected. Accepts an explicit bundle contract (trace_path + scenario_ids); scans for ready scenarios only when no contract is given.

Use when the problem is vague, the domain unfamiliar, multiple approaches are possible, or the user says 'investigate first' / 'before we spec this'. Read-only stance (not a workflow): no code changes, no mandatory artifacts; exits to spec/debug/clarify/direct implementation when insight crystallizes.

Use when the user gives a vague or high-level feature request that needs decomposition into implementable work — new ideas, '~만들어줘' with implicit requirements, features with multiple possible approaches. NOT for specific 1-2 file changes, obvious bug fixes, quick questions, or when spec/trace already exist (use stv:do-work).

Use when the trace backlog has ZERO unfinished scenarios (or the user explicitly asks for new feature ideas) and the project needs its next feature proposed. Hands the chosen idea to new-task for spec + trace.

Use when requirements and architecture need explicit definition before tracing — a new feature interview, or updating an existing spec.md (Update vs New). STV Phase 1: Proposal (WHY) -> Explore First (mandatory unknowns map) -> interview -> docs/{feature}/spec.md.

Use when the user wants real-world experience turned into a methodology, skill, or process — 'turn this into a methodology', 'systematize my approach', 'make this into a skill', or after stv:explore / completed work when raw practice should become reusable structure. Bottom-up inductive distillation, not top-down theory.

Use when the user says "resolve", "해결해줘", "픽스해줘", or wants to take one ready-to-agent llmux issue/PR, implement the fix on its branch, get `just check` green, and hand back a review-ready PR. Does NOT merge or deploy by default.

Use when the user says "triage", "이슈 정리", "분류", or wants to scan open llmux GitHub issues, decide which are safe for an agent to implement, label them, and open a tracking branch + draft PR for each ready one. Reads and labels issues; never writes source code.

Use for normal llmux dev iteration when the user says "빌드", "빌드해줘", "build", "build and deploy locally", or wants to compile + run a change locally and save it. Builds release, hot-deploys the binary to the local daemon, commits, and pushes to a feature branch (never main).

Use when the user says "배포", "배포해줘", "deploy", or "ship a preview" for llmux. Pushes to main, lets CI publish a preview prerelease, refreshes the llmux-preview brew formula, verifies brew actually updated, then hot-deploys that build locally and restarts.

Use when the user says "릴리즈", "릴리즈해줘", "release", "cut a release", or "stable release" for llmux. Bumps the version, tags v*, lets CI publish the stable GitHub release, refreshes the llmux stable brew formula, verifies brew updated, hot-deploys + restarts locally, and verifies client + server with llmux status.

Use when the user says "loop", "반복", "쭉 돌려줘", or wants to run triage then resolve repeatedly until there are no agent-workable llmux issues/PRs left. Thin orchestrator over the agent-triage and agent-resolve skills; dry-run by default.

Web scraping and data extraction using Scrapling. Use this skill whenever the user wants to scrape a website, crawl pages, extract data from URLs, search for products/prices online, compare prices across sites, or pull structured data from web pages. Also use when the user mentions Coupang, Amazon, shopping search, price comparison, or any task involving fetching and parsing web content — even if they don't explicitly say "scrape".

SNS thread scraping and timeline extraction skill for Twitter/X and Threads. Use this skill when the user wants to scrape someone's Twitter/X timeline, Threads posts, collect posts from a specific period, reconstruct threads (1/N, 2/N style), search for a person's recent posts on social media, or extract and organize social media content by date. Also trigger when the user mentions "트위터 긁어", "트윗 모아", "타임라인", "스레드 정리", "쓰레드 긁어", "threads.com", or any task involving fetching tweets, threads posts, or social media content.