도구

Unofficial CLI and MCP that let agents work with NotebookLM (now Gemini Notebook): bulk-import sources, ask questions with citations, and generate audio overviews, quizzes, flashcards, mind maps and slides. Uses undocumented APIs and may break; needs a Google sign-in and uv.

21st.dev's official MCP: lets agents search and fetch UI components, themes and template code from the 21st.dev community, and generate components from a description once AI access is enabled. Absorbs the former Magic MCP; requires a free API key from 21st.dev.

Veteran suite of CSV command-line tools: in2csv converts Excel and JSON to CSV, csvcut, csvgrep and csvsort filter and sort, csvstat computes statistics, and csvsql runs SQL directly against CSV files.

A prose linter that checks docs against style guides: use ready-made packages like Microsoft and Google or write your own YAML rules. It parses Markdown, AsciiDoc and other formats before checking, runs fully offline, and plugs into editors and CI.

Google's official skills for building ADK agents: dev workflow, scaffolding, the ADK Python API, evaluation with LLM judges and adaptive rubrics, deployment to Agent Runtime, Cloud Run or GKE, publishing and observability. Works with agents-cli; local dev needs only an AI Studio key.

Polishes existing motion against the transitions.dev token scale for duration, distance, scale, blur and easing, matching each value by what the motion does rather than the nearest number, with rules for open/close asymmetry, hover in and out, stagger and intent delays.

In-chat fallback for the transitions.dev Timeline Inspector's Refine button: the agent long-polls a local relay and answers jobs with motion-token tweaks or a full transition recipe swap, and applies accepted changes. Idle polling costs credits, so the persistent CLI mode is preferred.

Local neural text-to-speech engine that converts text to speech offline, with voices in dozens of languages, handy for listening materials or narrating articles and course slides. Needs Python; download a voice model before first use. The project is currently looking for maintainers.

Fixes defensive academic writing full of hedging and preemptive disclaimers, organizing abstract, introduction, experiments and conclusion around your strongest contribution. Also helps trim length and self-check before a rebuttal. English and Chinese versions, plus a prompt for any AI chat.

Posit's open-source scientific and technical publishing system built on Pandoc: embed Python, R or Julia code in Markdown or Jupyter notebooks, then quarto render produces HTML, PDF, Word, slides, websites and books.

Reference for VitePress 2, the Vite and Vue static site generator for docs: file routing, Markdown extensions and Vue in Markdown, data loaders, dynamic routes, theme config and custom themes, i18n, SSR compatibility and deployment. Generated by Anthony Fu from the official docs.

Semi-automated research workflow for computer science and AI researchers: ideation, Zotero literature management, experiment analysis and reports, paper writing and polishing, and rebuttals, with optional Obsidian knowledge base. 40+ skills; also installable as a Claude Code plugin.

Semi-automated research assistant for CS and AI researchers that links ideation, Zotero literature management, experiment analysis and reporting, paper writing and reviewer responses, and keeps project knowledge in Obsidian. Branches for Codex and other agents exist.

Design engineering plugin for product interfaces such as admin panels and dashboards: builds UI on design principles and saves spacing, color and hierarchy decisions to .interface-design/system.md to stay consistent across sessions. Not meant for marketing pages.

Connects to your local or cloud Zotero library: search by title, full text or semantics, read PDFs and annotations, add items by DOI and organize collections. Needs Python 3.10+; for local mode, enable the local API in Zotero settings first.

12 LinkedIn writing skills: write posts, comments and replies in your voice without AI tone, plan a week of content and rewrite your profile, waiting for your confirmation before each post. Produces copy-paste drafts by default; optional Apify (reading posts) and Publora (auto-publish) API keys.

12 LinkedIn writing skills: draft posts, comments and replies in your voice without obvious AI tone, optimize your profile and plan a week of content, with your approval before anything goes out. Also usable in Claude desktop/web (paid plans); optional Apify token to read posts, Publora to publish.

Lossy PNG compression CLI: converts images to 8-bit PNGs that keep transparency, usually 60-80% smaller, with batch processing and a quality range setting. Handy for slimming down web assets, slides and article images. Install via Homebrew on macOS; no account.

Lets agents generate production-ready Lottie animations (icon motion, loaders, data animations and more) from a text description or SVG, with a local player for live preview and tweaks. Export JSON for web, apps or After Effects. Needs Node.js locally.

High-quality GIF encoder built on pngquant: turns video frames or PNG sequences into animations, using cross-frame palettes and temporal dithering for thousands of colors per frame. Can take video piped straight from ffmpeg.

Audits and rewrites AI-sounding prose: detects stock phrases using 70+ pattern categories and a tiered replacement word list, with detect-only, rewrite and in-place edit modes and optional voice settings. Rules are calibrated for English.

Turns SRT subtitles into a hand-drawn whiteboard animation on warm paper that follows the narration, exported as MP4, pausing for approval at storyboard, line art and annotation steps. Good for lesson explainers. Needs Python; line art from an image-capable agent or your own.

From the FinanceToolkit author: computes ratios, valuation, risk and macro indicators with 500+ transparent formulas, for company comparisons and first-pass due diligence. Hosted, no install; enter a Financial Modeling Prep key on first connect (free tier: US stocks, 5 years). Verify results.

Detects shot changes in video and splits long footage into scenes, exporting cut lists and per-scene thumbnails, useful for shot analysis, organizing footage and finding edit points. Runs locally; needs Python 3.10+, plus FFmpeg or mkvmerge for splitting.