ツール

Teaches agents to code against the latest Gemini API and Python/TypeScript SDKs: multi-turn chat, function calling, structured output, and image and speech generation, while avoiding deprecated older models. Maintained by Google's Gemini team.

Spec-driven development MCP: the agent writes requirements, design and task docs in three steps, then implements task by task. A web dashboard or VS Code extension lets PMs review, comment, approve and track progress. Needs Node.js; launch the dashboard separately via npx.

Create, read and modify xlsx workbooks without Microsoft Excel installed: formulas, formatting and conditional formatting, charts, pivot tables and tables. Runs locally over stdio or as a remote HTTP service.

Keeps agents from reporting done too early on substantial work: writes acceptance gates with runnable checks before starting, runs only user-approved commands, splits large jobs into a depth tree of verified leaves, and re-checks evidence before reporting. Optional Claude Code Stop hook.

Writes or rewrites English docs in the spirit of ASD-STE100 Simplified Technical English: short sentences, active voice, condition before command, one word one meaning, terms defined at first use. Plain mode by default, Strict mode for STE compliance, plus a rule-by-rule check mode.

Open-source marketing plugin with 48 skills: SEO and AI search optimization (GEO), GA4 and Search Console analysis, ad diagnostics for Google, Meta, X, LinkedIn, Reddit and TikTok, and WordPress upkeep. Some SEO skills need only a URL; live data requires NotFair OAuth and linked accounts.

From Dagger: gives each coding agent its own container and git branch so multiple agents work in parallel without interfering, with command logs and the ability to take over the terminal at any time. Connects via MCP; needs Docker and Git. Still experimental.

From Dagger: gives every coding agent an isolated container and Git branch so several agents can work in parallel, and you can view command logs, take over the terminal or discard results anytime. Needs Docker and Git locally; still experimental.

Knownsec's open-source Python-use agent: describe a task in plain language and it writes and runs Python on your machine, e.g. tidying Excel files, summarizing attendance or screening resumes. Needs Python 3.11+ and a model API key such as DeepSeek; code runs locally, so back up important files.

Hugging Face's official hf CLI: search, download and upload models and datasets, manage repos and storage buckets, and run jobs on Hugging Face's cloud. Output adapts automatically for agents, and hf skills add installs companion skills. Uploads and similar actions need a login.

Expo's official plugin covering Expo Router navigation, native UI, data fetching and SDK upgrades, plus EAS builds, OTA updates and app store submission. Some service-related skills rely on paid EAS.

Three.js animation reference with code: procedural motion with Clock, the AnimationClip, AnimationMixer and AnimationAction keyframe system, playing GLTF animations, skeletal animation, morph targets and blending multiple animations.

Three.js fundamentals with code: setting up the scene, cameras and WebGLRenderer, Object3D hierarchy, coordinate systems and transforms, the render loop and resize handling. A starting point for building a 3D scene on the web.

Three.js lighting reference with code: ambient, hemisphere, directional, point, spot and area lights with their cost and shadow support, shadow setup, image-based environment lighting and lighting performance tips.

Three.js interaction reference with code: raycasting for click detection and object selection, camera controls such as OrbitControls, and mouse and touch input handling for interactive 3D experiences.

Three.js materials reference with code: when to use Basic, Lambert, Phong, Standard, Physical, Toon and other materials, PBR properties such as roughness and metalness, custom ShaderMaterial and material performance.

Three.js texture reference with code: loading and configuring texture maps, UV mapping, cubemaps and HDR environment maps, texture settings and optimization.

Three.js asset loading reference with code: loading GLTF models, textures, images and HDR environments, coordinating loaders and progress with LoadingManager, and async loading patterns.

Three.js post-processing reference with code: chaining passes with EffectComposer, adding bloom, depth of field, color grading, blur and glow, and writing custom screen-space shader passes.

Three.js geometry reference with code: parameters for built-in shapes like boxes, spheres, cylinders and tori, building custom meshes from vertices with BufferGeometry, and instanced rendering for large numbers of objects.

Three.js shader reference with code: writing GLSL vertex and fragment shaders, ShaderMaterial versus RawShaderMaterial, uniforms and animating them, and extending built-in materials for custom visual effects.

Open-source PDF table extraction with a camelot CLI: exports tables from financial reports, bank statements and statistical reports to Excel, CSV or Markdown, with an accuracy score. Needs Python 3.10+; handles text-based PDFs by default, scans need the ml and ocr extras. Double-check results.

Fast Rust toolkit for tabular data with nearly 80 subcommands for querying, stats, cleaning, validation, joins and format conversion on CSV and Excel. Also offers an MCP server and a Claude Cowork plugin.

NVIDIA's official, verified catalog of hundreds of skills covering CUDA-X libraries, NeMo training, NIM and Dynamo inference, RAG blueprints, Jetson, Omniverse and robotics simulation. The install adds only nvidia-skill-finder, which recommends other skills and installs them once you confirm.