Writes or improves a per-project skill explaining how to build, launch and drive the app.
Runs only when you type /run-skill-generator
Collects the hidden system prompts of popular AI chatbots and coding agents, updated daily.
Writes or improves a per-project skill explaining how to build, launch and drive the app.
Runs only when you type /run-skill-generator
Bundles a React design system and uploads it to claude.ai/design.
Runs only when you type /design-sync
Draws clear inline-SVG diagrams for Artifacts that stay legible in both themes.
Adds live data, saved state, viewer identity and device access to published Artifact pages.
Draws accurate maps from real geographic data for any deliverable that needs one.
Builds a three.js 3D model you can download as an OBJ or GLB file.
Answers questions about the Claude API: model ids, pricing, streaming, tool use, caching and migration.
Creates timeline-based motion design animations.
Gives design guidance and fundamentals for building Artifacts.
Learns your writing voice from your messages and docs so future drafts sound like you.
Researches topics using live web sources and returns findings grounded in them.
Builds a design system or UI kit when you ask for one.
Scans your transcripts for common read-only tool calls and adds an allowlist to reduce permission prompts.
Guides distinctive visual design for new or existing UI, covering aesthetic direction and typography.
Runs a design process for high-fidelity, polished interfaces before building them.
Exports the current work as a print-ready PDF file.
Explores many design ideas quickly through wireframes and storyboards.
Lets HTML prototypes call Claude directly through window.claude.complete.
Makes a single-page design ready for printing.
Produces a single-file HTML email that is ready to send.
Builds a working app prototype with real interactions.
Builds a slide presentation as an HTML file you can open and present.
Creates a page-style document that prints cleanly straight out of the box.
Adds on-canvas controls so you can adjust design values without editing code.
Shows several design options stacked vertically so you can compare and pick one.
Launches and drives the project's app so you can see a change working in the real app.
Saves the current work as one self-contained HTML file that works offline.
Creates, updates, lists or runs scheduled cloud agents that execute on a cron schedule.
Reviews the pending changes on the current branch for security problems.
Reviews changed code for reuse and simplification, then applies the cleanups.
Configures Claude Code settings, including hooks, permissions and environment variables.
Automates your Chrome browser to click, fill forms, take screenshots and read console logs.
Diagnoses Muse Code product and runtime issues from installed binary evidence when you explicitly ask.
Exports slides as native text and shapes that stay editable in PowerPoint.
Exports slides as flat images that look pixel-perfect but cannot be edited.
Packages the current work into a handoff bundle for a developer to pick up.
Creates a new CLAUDE.md file documenting your codebase for the agent.
Customizes keyboard shortcuts and chord bindings in your Claude keybindings file.
Runs a prompt or slash command again and again on a set interval you choose.
Plans a large-scale code change, then runs it in parallel across many isolated agents that each open a PR.
Runs only when you type /batch
Turns on debug logging for the session and helps diagnose problems.
Runs only when you type /debug
Creates new skills, edits existing ones and refines their wording for reliable triggering.
Runs a code change end-to-end to confirm it actually behaves as intended.
Runs only when you type /verify
Reviews a diff, pull request or branch for bugs at a chosen depth, and can post comments.
Imports a memory export from another AI assistant into Claude's memory.
Creates and edits shareable documents like reports, proposals and resumes, exporting to Word, PDF or Google Docs.
Creates, edits and converts Word documents, including revisions, comments and export to PDF or images.
Renders your morning brief as a styled page or sets it up as a weekday routine.
Makes PDFs from reports, posters, papers and resumes, and converts Office or HTML files to PDF.
Creates, edits, reads and converts PowerPoint decks, including exporting slides to PDF or images.
Opens, edits, creates and converts spreadsheet files like XLSX, CSV and TSV, including charts and data cleanup.
Guides creating or changing Google Docs, Sheets and Slides files in your Drive.
Keeps Markdown tables readable in a narrow terminal by reformatting them.
Applies visual design constraints before building or styling a web page or app.
Runs a goal as a project, coordinating parallel work threads and reporting their status.
Delivers a runnable local browser app you can operate right away.
Creates and validates a new native Muse plugin package in your workspace.
Creates and validates a new Muse skill, either project-local or personal.
Keeps a small focused test with the project when fixing bugs or changing behavior.
Gives one digest of coding-agent sessions across your machines and lets you open, steer and approve them.
Keeps Git operations safe by never committing, pushing or rewriting history unless you explicitly ask.
Scaffolds a brand-new project only when you explicitly ask to start one now.
Runs a decision interview you request and records each settled choice in project docs.
Runs a decision interview you request and records each settled choice in project docs.
Recovers read-only transcripts and resume checkpoints, or continues work from other coding agents.
Answers questions about and changes Muse Code settings such as model and reasoning effort.
Brings your saved memory notes and MCP servers from Claude Code or Codex into Muse Code.
Researches a request and returns one concise plan before any code is written.
Sets up Python environments inside the project directory and picks the right installer and run steps.
Finds and reads the agent's own past session logs to recover context or lost work.
Asks focused questions when a product decision would materially change scope, platform or architecture.
Continues unfinished Claude Code work inside Muse Code, with or without a session ID.
Continues unfinished Codex work inside Muse Code, with or without a session ID.
Helps you research and compare products while shopping online.
Explains how to use the browser pane inside the Claude desktop app before you start.
Creates, edits and comments on Word and Google Docs files, checking rendered pages before delivery.
Sends your feedback about Muse to the team from inside the agent.
Creates a shared live document that you and the agent can read and edit together.
Checks web pages for accessibility problems like labels, focus, keyboard use and contrast.
Explains how to use Claude in Chrome to act in your real browser with your own sign-ins.
Controls desktop apps on your computer by taking screenshots, clicking, typing and scrolling.
Opens, navigates, clicks and screenshots web pages and localhost targets inside ZCode.
Builds dashboards, KPI trackers and recurring reports with charts, numbers and tables.
Explains how to create and revise a design system artifact, including files, limits and checklist.
Removes stored information the agent remembers about you or the project.
Generates a podcast audio episode from your content.
Helps set and track goals for the work the agent is doing with you.
Reads and queries the Muse database for information the agent needs.
Turns text into spoken audio.
Debugs web pages, automates browser interactions and inspects performance and network requests.
Diagnoses and speeds up Largest Contentful Paint so your page's main content loads faster.
Finds and fixes problems with custom slash commands in the ZCode client.
Diagnoses and fixes ZCode hook configuration problems, such as hooks that never trigger or block sessions.
Diagnoses and fixes ZCode MCP server connection problems and missing tools.
Diagnoses and fixes ZCode plugin and marketplace problems, including install failures and missing skills.
Helps write, debug and resubmit dynamic-workflow scripts that coordinate subagents for the CreateWorkflow tool.
Renders interactive HTML widgets inline in chat, such as diagrams, charts and controls.
Creates or updates ZCode plugin source and a local test marketplace, then guides installing and trying it.
Creates a short motion animation film.
Creates interactive visualizations, simulations, maps and mockups in the conversation to explain how something works.
Explains where and how to configure ZCode's MCP servers, commands, skills, hooks, plugins and AGENTS.md files.
Explains how Antigravity customizations load and helps you create skills, rules, plugins and hooks.
Answers questions about using and configuring Google Antigravity, its CLI, IDE and SDK.
Guides you through setting up a recurring scheduled task, like a daily email summary.
Writes code that calls the Gemini API for text, chat, images, video, speech and structured output.
Builds real-time streaming apps with the Gemini Live API, including audio, video and voice detection.
Generates and edits video with Gemini Omni Flash, including text-to-video and frame transitions.
Finds and fixes memory leaks in JavaScript and Node.js apps using heap snapshots.
Converts old workflow files into modern skills and archives the originals safely.
Installs, enables, disables or removes plugins, and packages customizations into a single installable bundle.
Diagnoses browser connection and page-target failures using Chrome DevTools.
Merges duplicate memory entries, fixes stale facts and prunes the memory index.
Explains where this session's tokens went, with a simple chart in plain language.
Guides setup by installing matching plugins, connecting your tools and trying a skill.
Generates a report analyzing your Claude Code sessions.
Runs only when you type /insights
Sends the work from the current session into a Claude Project.
Runs only when you type /send-to-project
Creates a session-start hook committed to your repository for Claude Code cloud sessions.
Sets up the status line shown at the bottom of Claude Code.
Runs only when you type /statusline
Writes an onboarding guide that helps teammates get started with Claude Code from your usage.
Runs only when you type /team-onboarding
Picks the right tool to read an uploaded file by type, such as PDF, Word, spreadsheet or image.
Reads and extracts text, images, tables and form fields from PDF files.
Answers specific questions about Claude Code, the Claude API and Claude.ai plans, limits and setup.
Checks whether datasets and query results are trustworthy, flagging freshness, duplicates and broken joins.
Builds an analytics dashboard spreadsheet tracking acquisition, retention, revenue and funnel KPIs.
Creates a business review presentation covering KPIs, segment results, priorities and outlook.
Writes a design report with executive summary, findings, recommendations and appendix.
Writes an experiment analysis document with hypothesis, method, results, limits and next steps.
Builds a financial budget spreadsheet with actuals, forecasts, variances and cash runway.
Writes an investment committee memo with thesis, deal details, financials, risks and recommendation.
Builds or updates an interactive dashboard from connected data, spreadsheets or CSV files.
Writes polished analytical reports for executive, product or technical readers, backed by evidence.
Turns an existing Data app into a polished DOCX or Google Doc.
Turns an existing Data app into a polished PowerPoint or Google Slides deck.
Saves reusable analysis settings, conventions and data definitions for future reports and dashboards.
Guides chart and dashboard design, including colors, stat tiles and layout, before you write chart code.
Runs many web searches, checks sources, verifies claims and writes a cited report.
Runs only when you type /deep-research
Defines success metrics, targets, guardrails and measurement plans for product decisions.
Answers questions about OpenAI DevDay attendance, schedules, sessions and recordings.
Suggests working with data from your devices, based on the skill name.
Diagnoses and fixes ZCode skill problems, such as skills not loading or not triggering automatically.
Gathers business context from connected sources so analysis starts with the right framing.
Creates and edits native Google Docs through your connected Google account.
Finds, organizes, shares and edits Drive files, Docs, Sheets and Slides from one place.
Writes, replies to and resolves Google Drive comments on Docs, Sheets, Slides and files.
Analyzes and edits Google Sheets with precise cell ranges, formulas and charts.
Edits, restyles and repairs existing Google Slides decks using a reference template's design.
Walks you through using the agent's features and how to get started.
Searches for images to use in your work.
Answers business and product questions with data and routes work to the right data workflow.
Creates, edits and validates reproducible SQL or Python notebooks you can rerun.
Prepares KPI readouts, scorecards and executive summaries from business metrics.
Estimates market, segment or opportunity size with clear assumptions and stated uncertainty.
Manages a library of media files the agent can browse and reuse.
Explains why a metric moved, dropped or differs from what you expected.
Handles signing up for or checking early access to Muse features.
Builds UI components outside Storybook, without its preview tooling.
Opens Codex Security Cloud to view security findings, scans and repository monitoring.
Diagnoses failed OpenAI API requests and points you to the right fix.
Sets up and safely handles your OpenAI API key when building AI-powered apps.
Analyzes product or business data to back a decision with metric-based evidence and recommendations.
Publishes an existing data report or dashboard to Sites so it is available on the web.
Turns an existing data dashboard or report into a polished PDF file.
Sets up recurring cloud jobs that keep an existing data dashboard or report up to date.
Lets the agent check and describe its own capabilities and configuration when asked.
Guides Cowork setup by installing a matching plugin, trying a skill and connecting tools.
Works with spreadsheet files to create, read and edit their contents.
Works with Storybook to build and preview UI components in isolation.
Checks your subscription status and plan details.
Helps write and run tests for your code.
Provides Three.js patterns and API guidance for building 3D scenes.
Helps you plan a trip, including destinations, timing and logistics.
Checks an analysis's sources, calculations, visuals and conclusions for accuracy and completeness.
Designs, builds and verifies quantitative charts and figures for reports, dashboards and notebooks.
Handles communication with wearable devices, based on what the name suggests.
Tests web frontends like a real user by clicking, typing and scrolling in a browser, then reports results.
Creates and edits shared whiteboards for laying out thinking in whatever form it needs.
Runs broad research across many sources, based on what the name suggests.