도구

macOS screen capture, accessibility-tree inspection and native UI automation, served over MCP so agents can see the screen and click, type and operate menus. Requires macOS 15+ with Screen Recording and Accessibility permissions granted.

macOS screenshot and UI automation CLI: capture the screen, read the accessibility tree for structured UI elements, then click, type, and switch windows and menus. Also runs as an MCP server for Claude Code and Codex. Needs macOS 15+ with Screen Recording and Accessibility permissions.

LM Studio's official CLI: list, load and unload local models, start and stop the local API server, and check status and logs; ls, ps and other commands support --json output for scripts and agents. Requires the LM Studio desktop app.

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.

Lingo.dev's official CLI: lingo push sends JSON, YAML, Markdown, PO, XLIFF, Android/Xcode strings and other files from your repo to its LLM localization engine, and lingo pull brings translations back; CI-ready. Needs a Lingo.dev account; usage-based pricing with a free sandbox.

Mistral's official open-source terminal coding assistant: uses Mistral models to read and write code, run commands and search the codebase, with multiple agent profiles and subagents, plus Skills, MCP and a --prompt scripting mode. Needs a Mistral API key.

From Callstack: lets coding agents drive and verify iOS, Android, HarmonyOS and desktop apps through accessibility snapshots, with taps, form filling, screenshots, logs and performance sampling, plus a built-in MCP server. Requires Node.js 22.12+.

From Callstack: lets coding agents inspect, operate and verify apps on iOS, Android and HarmonyOS simulators and real devices, reading accessibility snapshots, taking screenshots and recordings, and capturing logs and performance data. Ships both a CLI and a built-in MCP server.

Font toolkit: pyftsubset subsets fonts to the characters you use and outputs WOFF2 to shrink web fonts, and ttx converts fonts to editable XML for inspection and edits. Needs Python 3.11+ and the woff extra for WOFF2; check the license before subsetting commercial fonts.

Classic CLI for reading and writing metadata in hundreds of file types including images, video, audio and PDF, covering EXIF, GPS, IPTC, XMP and camera makers' proprietary tags, with batch editing, stripping and conversion. Requires Perl.

Generates hand-drawn-style Excalidraw diagrams (.excalidraw files you can keep editing on excalidraw.com) from text, choosing fan-out, timeline or convergence layouts by how concepts relate. Optional uv and Playwright let the agent render and check layouts; colors live in one file you can restyle.

Converts AsciiDoc, a Markdown-like plain-text markup, into HTML5, DocBook and man pages, and with official extensions into PDF and EPUB, suited to technical books, textbooks and large manuals. Ruby-based; install with Homebrew on macOS.

Gives agents Exa web search and page content fetching, with an optional multi-step research agent and advanced filtered search. Maintained by Exa; the hosted service works anonymously with limits, and signing in or adding an API key raises them.

WordPress's official CLI: install and update plugins and themes, manage posts and users, import and export content, run database search-replace and configure multisite from the terminal, without opening wp-admin. Requires PHP 7.2.24 or later.

CLI and TypeScript runtime for discovering and calling MCP servers from the terminal and scripts: list tool signatures and call tools directly, or generate a standalone CLI or typed client from a server. Auto-imports existing MCP configs from Claude Code, Codex and others.

Reads any website's design system in a headless browser and exports DTCG design tokens, Tailwind config, Figma variables and a brand guide in one command, scoring contrast and cross-page consistency. Also an MCP server or Claude Code plugin. Needs Node.js and Chromium or Chrome.

Design token build tool: define colors, font sizes, spacing and other design specs in one JSON source, then style-dictionary build generates CSS, SCSS, iOS, Android and other platform code, keeping the design system and implementations in sync. Needs Node.js and a config.json.

Read-only mirror of the community plugin marketplace maintained by Anthropic, with 2,000+ community plugins that passed submission review and automated security scans, for Claude Code and Claude Cowork.

Cloudflare's official Workers CLI: develop and debug Workers locally, deploy them, manage D1, R2, KV and other resources, and tail live logs. Cloudflare recommends installing it per project; requires a Cloudflare account.

Moves files to the trash instead of deleting them, with commands to list, restore and empty trashed files, so anything an agent deletes by mistake can be recovered. Follows the FreeDesktop.org trash spec and shares the trash with GNOME and KDE.

Stores agent memory as local Markdown notes that both people and AI can read, write, search and link, and you can open them directly in Obsidian. Local-first, open source and free, with a paid cloud sync option.

Firebase's official CLI: deploy Hosting, Cloud Functions and security rules, run local emulators, read and write databases and import or export users. The same repo also provides the official Firebase MCP server. Requires a Firebase sign-in.

Lets agents manage Firebase projects: read and write Firestore and Realtime Database, manage Auth users, validate security rules, and check Crashlytics crashes and function logs. Maintained by Firebase; sign in to the Firebase CLI first.

Teaches agents to write video prompts for ByteDance's Seedance 2.0: input limits, @ asset references and camera language, with templates for product ads, short dramas, dance, scenic clips, e-commerce and explainers, in English and Chinese. Prompts only; generate videos with your own account.