agenthub
agenthub contains 5 collected skills from jeongsk, with repository-level occupation coverage and site-owned skill detail pages.
Skills in this repository
YouTube ์์์ ๋ค์ด๋ก๋ํด ์๋ง๊ณผ ์คํฌ๋ฆฐ์ท์ ๊ฒฐํฉํ ํ๊ตญ์ด ํ์ต ๋ ธํธ(.md)๋ก ์ ๋ฆฌํ๋ค. "YouTube ์์ ์ ๋ฆฌ", "์์ ํ์ต๋ ธํธ", "์ ํ๋ธ ๋ ธํธ", "์๋ง ๊ธฐ๋ฐ ๋ ธํธ", "์์ ์์ฝ ๋งํฌ๋ค์ด", "video to notes", "YouTube summary in Korean", "study notes from video", "transcribe and screenshot" ๊ฐ์ ํํ์ ๋ฌผ๋ก , ์ฌ์ฉ์๊ฐ YouTube/์ ํ๋ธ/youtu.be ๋งํฌ๋ฅผ ๋์ง๋ฉฐ "์ ๋ฆฌ", "์์ฝ", "๊ณต๋ถ ์๋ฃ", "์คํฌ๋ฆฝํธ", "์บก์ฒ์ ๊ฐ์ด" ๋ฑ์ ์ธ๊ธํ๋ฉด ๋ฐ๋์ ์ด ์คํฌ์ ์ฌ์ฉํ๋ค. yt-dlp๋ก ์์ยท์๋ง์ ๋ฐ๊ณ , ffmpeg๋ก ํต์ฌ ์ฅ๋ฉด์ ์บก์ฒํ๊ณ , ์คํฌ๋ฆฐ์ท-์๋ง ๋งค์นญ์ ๋ชจ๋ธ์ด ์ง์ ๊ฒํ ํ ๋ค ์์์ ๋ณด์ง ์์๋ ์ดํด๋๋ ํ๊ตญ์ด ๋ ธํธ๋ฅผ ๋ง๋ ๋ค.
์ ์ฅ์์ ์ด๋ฏธ์ง๋ฅผ ์น์ฉ์ผ๋ก ์ต์ ํํ๋ค(WebP ๋ณํยทํด์๋ ์ถ์ยทGIF ์ฒ๋ฆฌ). PNG/JPG/๋ํ WebP๋ฅผ WebP๋ก ์ฌ์ธ์ฝ๋ฉํ๊ณ , ํ๋ฉด ์บก์ฒํ ๋ํ GIF๋ ์ ์ ํ๋ ์์ผ๋ก, ์งง์ ๋ฐ๋ชจ GIF๋ ์ ๋๋ฉ์ด์ WebP๋ก ์ค์ธ ๋ค ๋ณธ๋ฌธ์ ์ด๋ฏธ์ง ์ฐธ์กฐ๊น์ง ์ค์ ํ์ผ๋ช ๊ณผ ๋ง์ถ๋ค. ์ฌ์ฉ์๊ฐ "์ด๋ฏธ์ง ์ต์ ํ/์ฉ๋ ์ค์ฌ/์คํฌ๋ฆฐ์ท ์ต์ ํ/optimize images"๋ฅผ ์์ฒญํ๊ฑฐ๋, public/tool-screenshots ๊ฐ์ ์ ์ ์ด๋ฏธ์ง ๋๋ ํฐ๋ฆฌ๊ฐ ์ปค์ก์ ๋ ์ฌ์ฉํ๋ค.
Maintain a live implementation-notes.md while implementing a feature from a SPEC. Records ambiguous design decisions, intentional deviations, tradeoffs, and open questions as they happen so the user can give feedback during implementation, not after. Use after grill-me, or whenever the user asks to implement a SPEC / specification / design doc / requirements. Triggers include "์คํ ๊ตฌํ", "specification ๋ฐ๋ผ ๊ตฌํ", "design doc ๋ฐ๋ผ ๊ตฌํํด์ค", "implement this spec", "build this from the spec".
Research a user-requested skill, plugin, MCP server, or browser extension and add a verified Markdown entry to AgentHub's content registry. Use when a user asks to register, submit, add, or list a new registry item.
Generate a Pinterest/Instagram-style analog Korean reading journal image from a book title, author, finish date, rating, and optional notes. Use when the user wants a one-page handwritten scrapbook-style book memo image. Ask for missing required information first, then create the image directly with the available image generation capability.