__Vault

The master coordinator for AI skills. Discovers skills from multiple sources (SkillsMP.com, SkillHub, and ClawHub), manages installation, and synchronization across Claude Code, Gemini CLI, Google Anti-Gravity, OpenCode, and other AI tools. Handles User-level (Global) and Project-level (Local) scopes.

Reviews UI code, components, and HTML for WCAG 2.1 accessibility issues and produces a prioritized remediation list with fixes. Use this skill whenever a user says "check this for accessibility", "is this component accessible?", "WCAG audit this code", "fix the a11y issues here", "review for screen reader support", "check color contrast", "make this keyboard navigable", or "audit this form for accessibility". Also activate when someone says "my accessibility test is failing" or pastes HTML/JSX and asks if it's accessible. Do NOT use for full automated accessibility testing tool setup (axe-core, Lighthouse) or legal ADA compliance certification.

Writes CSS keyframe animations, transitions, and JavaScript-driven motion effects from natural language descriptions. Use this skill whenever a user says "animate this element", "write a CSS animation for...", "make this fade in", "add a slide transition", "create a loading spinner animation", "animate on scroll", "write a Framer Motion animation", or "how do I make this bounce/pulse/spin?". Also activate when someone describes a UI interaction that should feel smooth or delightful. Supports CSS animations, CSS transitions, Web Animations API, and Framer Motion. Do NOT use for complex 3D animations (Three.js/WebGL), game animations, or SVG path morphing.

Scaffolds API client code from an OpenAPI/Swagger spec, a list of endpoints, or a description of an API. Generates typed client functions, error handling, retry logic, and authentication setup. Use this skill whenever a user says "write an API client for this", "scaffold a client from this OpenAPI spec", "generate API wrapper functions", "help me integrate with this API", "write the fetch calls for these endpoints", "create an SDK wrapper", or "give me typed API functions for this service". Also activate when a user shares an API spec or endpoint list and needs code to call it. Supports Python, TypeScript, and JavaScript. Do NOT use for designing new APIs (use api-design-skill) or writing backend route handlers.

Generates architecture diagrams as Mermaid code from system descriptions, service lists, or existing documentation. Supports flowcharts, sequence diagrams, entity-relationship diagrams, and C4 context diagrams. Use this skill whenever a user says "draw an architecture diagram", "create a Mermaid diagram for this", "diagram this system", "show me a sequence diagram for this flow", "generate an ERD for these tables", "visualize this API flow", "draw the data flow for this", or "make a diagram of how these services connect". Also activate when a user describes a system and wants it visualized. Do NOT use for creating actual architecture (use system-design approaches), generating image files directly, or creating UI wireframes.

Designs and scaffolds background job, queue worker, and task scheduling patterns: job definitions, queue configuration, retry policies, dead-letter handling, and monitoring hooks. Use this skill whenever a user says "add a background job for...", "process this asynchronously", "schedule this task", "write a queue worker", "set up a job queue", "add retry logic for this operation", "process emails in the background", "implement a job system", or "how do I run this without blocking the request". Also activate when a user has a slow or unreliable operation that shouldn't block an API response. Supports BullMQ/Node, Celery/Python, and Go goroutine patterns. Do NOT use for real-time event streaming (use Kafka/event-sourcing patterns) or scheduled cron infrastructure setup.

Recommends and implements caching strategies for APIs, databases, and frontend assets: cache layers, TTLs, cache invalidation patterns, and Redis/in-memory implementation code. Use this skill whenever a user says "add caching to this", "my API is too slow — should I cache?", "implement Redis caching", "what's the right TTL for this?", "how do I cache this database query?", "add HTTP caching headers", "implement a cache-aside pattern", or "optimize performance with caching". Also activate when someone has a slow endpoint or repeated expensive query. Supports Redis, in-memory (Node, Python), and HTTP cache headers. Do NOT use for CDN configuration (infrastructure-level) or browser service worker caching (different skill scope).

Writes structured, human-readable changelogs from commit histories, PR lists, or raw notes. Use this skill whenever a user says "write a changelog", "generate CHANGELOG.md", "summarize what changed in this release", "turn these commits into release notes", "write a changelog for v2.3", or "what changed since last version?" Also activate when someone pastes a list of commits or PRs and asks what to put in a changelog. Follows Keep a Changelog conventions by default. Do NOT use for writing new code, creating PR descriptions, or drafting marketing copy for releases.

Scaffolds command-line interface (CLI) tools with subcommands, flags, help text, input validation, and colored output. Use this skill whenever a user says "build a CLI for...", "create a command-line tool that...", "write a CLI with subcommands", "scaffold a Node CLI", "make a Python CLI with argparse or Click", "write a Go CLI tool", or "create a dev tool I can run from the terminal". Also activate when someone describes a workflow they run manually and wants it wrapped in a nice CLI. Supports Node.js (Commander.js / Yargs), Python (Click / argparse), and Go (cobra). Do NOT use for shell scripts (use python-script-generator instead) or full TUI (terminal UI) applications.

Analyzes code and applies structured refactoring improvements: extracts functions, reduces complexity, eliminates duplication, improves naming, and modernizes syntax. Use this skill whenever a user says "refactor this code", "clean up this function", "this function is too long", "reduce complexity here", "extract this into a helper", "make this more readable", "improve this code structure", or "simplify this logic". Also activate when someone shares code they feel is messy or hard to maintain. Supports JavaScript, TypeScript, Python, Go, Java, and Ruby. Do NOT use for fixing bugs (the code's logic must stay correct), writing new features, or performance optimization (use sql-query-optimizer or a profiling tool instead).

Generates a developer onboarding guide for a codebase: explains architecture, key directories, setup steps, coding conventions, and how to make a first contribution. Use this skill whenever a user says "write an onboarding guide for this repo", "help new devs get started with this codebase", "create a developer README", "write a contributing guide", "explain how this project is structured", or "help me document how to set up this project". Also activate when a new engineer is starting and needs a structured introduction to a codebase. Do NOT use for writing API reference docs or full system architecture documents.

Generates accessible, harmonious color palettes with contrast-checked values, semantic roles (primary, surface, error, etc.), and CSS/Tailwind token outputs. Use this skill whenever a user says "generate a color palette", "create a color scheme for my app", "pick colors for my design system", "make accessible colors", "generate color tokens", "what colors should I use for this brand?", "check if these colors pass WCAG", or "create a dark mode color palette". Also activate when someone describes a brand or visual mood and wants a concrete color set. Do NOT use for choosing fonts (typography skill) or icon/illustration color matching (art direction skill).

Creates CSS and Tailwind layout solutions for described UI patterns: grids, flex containers, sidebar layouts, card grids, sticky headers, centered content, and full-bleed sections. Use this skill whenever a user says "create a CSS layout for", "help me center this", "write CSS for a sidebar layout", "make a responsive grid", "how do I do a sticky header with scrollable content", "layout this page with CSS", "write Tailwind classes for this layout", or "fix my flexbox/grid layout". Also activate when someone shares a layout description or rough wireframe sketch. Supports plain CSS, CSS Grid, Flexbox, and Tailwind CSS v3/v4. Do NOT use for component-level styling (use react-component-generator) or animation (use animation-snippet-writer).

Transforms CSV data into structured, readable reports with summaries, key metrics, tables, and narrative insights. Use this skill whenever a user says "turn this CSV into a report", "generate a report from this data", "summarize this spreadsheet", "create a data report from these numbers", "write a business report from this CSV", "analyze this data and produce a report", or "convert this data into something readable". Also activate when a user shares tabular data and wants it turned into a polished document. Do NOT use for building interactive dashboards (use dashboard-design-skill) or writing full data cleaning pipelines.

Drafts a daily standup update from your recent work activity, commit history, calendar events, or bullet notes. Use this skill whenever a user says "write my standup", "draft my daily update", "what should I say at standup today?", "help me summarize what I worked on yesterday", or "write my async standup post". Also activate when someone shares recent commits, tickets, or a quick brain dump and asks for a standup-ready summary. Do NOT use for weekly reviews, meeting agendas, or project status reports (those are separate skills).

Adds dark mode and theme switching to existing UI codebases: defines dark-mode color tokens, writes CSS custom property overrides, implements a theme toggle component, and respects system preference. Use this skill whenever a user says "add dark mode to my app", "implement theme switching", "add a light/dark toggle", "make my app respect prefers-color-scheme", "write dark mode CSS", "implement a dark theme for this component", or "how do I add dark mode to my React app". Also activate when someone asks how to persist a user's theme preference. Works with plain CSS, Tailwind CSS, and React. Do NOT use for generating the color palette itself (use color-palette-generator first) or for generating full design system tokens.

Writes step-by-step data cleaning recipes using Python and pandas (or polars): handles missing values, deduplication, type coercion, outlier removal, string normalization, and format standardization. Use this skill whenever a user says "clean this data", "write pandas code to fix this dataset", "how do I handle missing values in this CSV?", "deduplicate this dataframe", "normalize these strings", "fix these date formats", "write a data cleaning pipeline", or "prepare this data for analysis". Also activate when a user shares a sample of messy data and asks how to fix it. Do NOT use for SQL-based data cleaning (use sql-migration-drafter) or full ETL pipeline design.

Designs and scaffolds ETL/ELT data pipeline architectures: data flow diagrams, extraction scripts, transformation steps, load patterns, and orchestration configuration. Use this skill whenever a user says "design a data pipeline", "build an ETL pipeline", "move data from X to Y", "extract data from this source and load into my warehouse", "schedule a data sync", "build a data ingestion pipeline", "design an ELT workflow", or "how do I automate data movement between systems". Also activate when someone describes a manual data movement process they want to automate. Supports Python (pandas, dbt, Airflow), and SQL-based patterns. Do NOT use for real-time streaming pipelines (Kafka, Flink) or for single data cleaning tasks (use data-cleaning-recipe-writer).

Audits project dependencies for outdated versions, known vulnerabilities, license conflicts, and unused packages. Use this skill whenever a user says "audit my dependencies", "check for vulnerable packages", "find outdated npm packages", "which dependencies should I update?", "are there any security issues in my package.json?", "check my requirements.txt for CVEs", or "help me clean up unused dependencies". Also activate when a user shares a package.json, requirements.txt, go.mod, or similar file and asks what to do with it. Do NOT use for writing new features, refactoring code, or general security audits of application logic.

Exports design system tokens (colors, spacing, typography, shadows, radii) to CSS custom properties, JavaScript/TypeScript constants, Figma tokens (W3C format), or Tailwind config extensions. Use this skill whenever a user says "export my design tokens", "convert these tokens to CSS variables", "generate a Tailwind config from my tokens", "create a JS tokens file", "write design tokens for my design system", "export tokens from Figma format", or "sync my design and code tokens". Also activate when someone wants to centralize colors, spacing, or typography across design and code. Do NOT use for generating the colors or spacing values themselves (use color-palette-generator first).

Rewrites and restructures existing documentation to a consistent standard: clearer headings, consistent style, proper Markdown formatting, and logical flow. Use this skill whenever a user says "clean up this doc", "rewrite this README", "make this documentation consistent", "improve this guide", "refactor these docs", "fix the formatting in this Markdown file", or "standardize our documentation style". Also activate when someone pastes a messy or outdated doc and asks for it to be improved. Do NOT use for writing new documentation from scratch (use documentation-generator) or for editing code, only prose docs.

Writes consistent, production-ready error handling patterns for a given language and framework: custom error classes, try/catch blocks, centralized error middleware, and structured error responses. Use this skill whenever a user says "write error handling for this", "add try/catch here", "create a custom error class", "how should I handle errors in this API?", "write an error middleware", "standardize my error responses", or "add proper error handling to this function". Also activate when code is missing error handling or has inconsistent error patterns. Supports Node.js/Express, Python/FastAPI, Go, and Java/Spring. Do NOT use for logging infrastructure setup (separate concern) or writing retry logic (use background-job-designer).

Builds evaluation (eval) sets for AI skills and prompts: generates realistic test prompts, defines expected outputs, and structures them into evals.json files. Use this skill whenever a user says "create evals for this skill", "build test cases for my prompt", "write an eval set", "how do I test this skill?", "generate prompts to test my AI feature", or "help me create a benchmark for this workflow". Also activate when a user has a skill or prompt and wants to know if it works correctly. Do NOT use for running evals (use eval-runner) or for creating the skills themselves (use skill-creator).

Builds HTML and React forms with field validation, error states, loading states, and accessible markup. Use this skill whenever a user says "build a form for...", "create a contact form", "write a registration form", "add validation to this form", "build a multi-step form", "handle form submission", "add form error handling", or "write a form with React Hook Form". Also activate when someone describes a set of fields they need to collect from a user. Works with plain HTML, React + React Hook Form, and React + Zod validation. Do NOT use for backend form processing/storage (use rest-endpoint-generator) or payment forms (Stripe Elements are a different integration).

Writes GraphQL schemas (SDL), resolvers, mutations, subscriptions, and input type definitions. Use this skill whenever a user says "write a GraphQL schema", "add a GraphQL mutation for...", "create a GraphQL type for...", "write a resolver for this query", "scaffold a GraphQL API", "add a subscription to this schema", "define GraphQL input types for...", or "convert my REST API to GraphQL". Also activate when someone describes a data model and wants it expressed as GraphQL types. Works with Apollo Server, graphql-js, and Pothos (code-first). Do NOT use for REST API endpoints (use rest-endpoint-generator) or full schema federation architecture (too complex for one skill).

Scaffolds an SVG icon component system: a reusable Icon wrapper component, a typed icon name union, an icon sprite approach or inline SVG approach, and a usage guide. Use this skill whenever a user says "build an icon component", "create an icon system", "make my SVGs reusable as components", "scaffold an icon library in React", "convert these SVGs to components", "how do I manage icons in my design system?", or "create a typed icon set". Also activate when someone wants to standardize how icons are used across their app. Works with React + TypeScript, Vue 3, or plain Web Components. Do NOT use for creating or sourcing the actual SVG icon artwork (use an icon library like Lucide, Phosphor, or Heroicons as the source).

Writes a structured incident postmortem (post-incident review) from raw notes, timelines, and team input. Use this skill whenever a user says "write a postmortem", "document this incident", "write an incident report", "help us do a blameless retrospective on this outage", "turn these incident notes into a PIR", "write a post-incident review", or "document what happened during the outage". Also activate when someone shares a timeline of events and asks for a structured incident document. Do NOT use for live incident response, writing runbooks for future incidents, or general meeting retrospectives.

Writes data transformation scripts to reshape JSON from one schema to another: field mapping, renaming, flattening nested objects, aggregating arrays, and producing clean output structures. Use this skill whenever a user says "transform this JSON", "reshape this data", "map these fields to a new format", "flatten this nested JSON", "convert this API response to my data model", "write a JSON transformer", "reformat this data structure", or "extract and remap these fields". Also activate when someone shows two JSON structures (source and target) and wants the mapping code. Works with JavaScript/TypeScript, Python, and jq. Do NOT use for parsing JSON from strings (use JSON.parse / json.loads) or for schema validation (use type-definition-generator + Zod).

Converts raw notes, brain dumps, documents, and unstructured text into structured knowledge base entries suitable for Notion, Confluence, a wiki, or a personal PKM system. Use this skill whenever a user says "turn these notes into a KB article", "add this to our knowledge base", "structure these raw notes", "write a wiki page from this", "convert this into a how-to guide", "make this searchable", or "ingest this into our docs". Also activate when someone pastes a long unstructured document and wants it organized. Do NOT use for meeting notes (use meeting-notes-to-actions) or code documentation generation.

Triages, summarizes, and categorizes error logs to identify root causes, patterns, and critical issues. Use this skill whenever a user says "help me read these logs", "triage these errors", "what's wrong in my logs?", "summarize these stack traces", "find patterns in these errors", "which errors are most critical?", "parse this log file", or "what's causing all these 500 errors?". Also activate when a user pastes a wall of log output and asks what to do. Works with application logs, server logs, CI logs, and database logs. Do NOT use for writing log collection infrastructure, log retention policies, or live monitoring setup.

Converts rough meeting notes, transcripts, or voice-memo summaries into structured action items, decisions, and follow-ups. Use this skill whenever a user says "turn these meeting notes into action items", "extract the todos from this call", "summarize this transcript", "who owns what from this meeting?", "write up the decisions from our standup", or "clean up these notes". Also activate when a user pastes a messy wall of text from a meeting and asks what to do next. Do NOT use for scheduling meetings, writing agendas for future meetings, or live transcription.

Organizes, tags, deduplicates, and improves a prompt library. Use this skill whenever a user says "organize my prompts", "tag these prompts", "clean up my prompt library", "find duplicate prompts", "sort my prompts by category", "write better versions of these prompts", "add metadata to my prompts", or "build a searchable prompt collection". Also activate when a user pastes a list of prompts and wants them structured or improved. Do NOT use for running prompts (that's a different operation) or for creating brand-new prompts from scratch without existing material to work with.

Improves existing prompts for better AI output quality: clearer instructions, better output format specification, role/persona addition, few-shot examples, and chain-of-thought encouragement. Use this skill whenever a user says "improve this prompt", "why is this prompt not working well?", "make this prompt better", "optimize this prompt for Claude/GPT", "my prompt gives bad results", "rewrite this prompt", "add examples to this prompt", or "help me get better outputs from this prompt". Also activate when a user shares a prompt and seems frustrated with the AI's outputs. Do NOT use for building full prompt libraries (use prompt-library-curator) or creating skills (use skill-creator).

Generates small, focused Python scripts for automation tasks: file processing, data transformation, API calls, CLI utilities, and task automation. Use this skill whenever a user says "write a Python script to...", "create a Python utility for...", "automate this with Python", "write me a script that...", "generate a Python CLI for...", or "help me write a Python program to...". Also activate when a user describes a repetitive task and wants it automated with Python. Always includes error handling, logging, and confirmation prompts for destructive actions. Do NOT use for full application development, data science notebooks, or machine learning model training scripts.

Implements rate limiting middleware with configurable rules, sliding window or token bucket algorithms, Redis-backed distributed limiting, and response headers. Use this skill whenever a user says "add rate limiting", "protect this endpoint from abuse", "implement a rate limiter", "limit API requests per user", "throttle this endpoint", "add rate limits to my API", "prevent brute force on login", or "implement per-IP and per-user rate limits". Also activate when someone's API is being hammered or abused. Supports Express/Node.js and FastAPI/Python with Redis (distributed) or in-memory (single-server). Do NOT use for CDN-level rate limiting (Cloudflare, AWS WAF — infrastructure scope) or queue throttling (use background-job-designer).

Generates React components with TypeScript props types, hooks, accessibility attributes, and basic test stubs. Use this skill whenever a user says "create a React component for...", "write a React component that...", "build a [name] component in React", "generate a reusable React component", "scaffold a React hook for...", or "add a component to my React app". Also activate when someone describes UI behavior and wants it as a React component. Produces modern functional components with hooks; does not generate class components. Works with React 18+, TypeScript, and either CSS Modules, Tailwind, or styled-components. Do NOT use for full-page layout scaffolding (use css-layout-helper) or state management architecture.

Builds, explains, and debugs regular expressions from natural language descriptions. Use this skill whenever a user says "write a regex for...", "build a regular expression that matches...", "explain what this regex does", "why isn't my regex matching?", "help me write a pattern for...", "validate this format with regex", or "match these strings but not those". Also activate when someone shares a regex they don't understand or one that isn't working. Works for all major flavors: JavaScript, Python (re), PCRE, Go, Ruby. Do NOT use for full text parsing of structured formats like JSON/XML (use a proper parser instead) or for fuzzy text matching.

Generates polished, user-facing release notes for product launches, version bumps, and feature announcements. Use this skill whenever a user says "write release notes", "draft an announcement for this release", "generate patch notes", "what should I tell users about this update?", or "write a product update email". Also activate when a user wants release copy for app store listings, blog posts, or in-app banners. This skill focuses on clear, benefit-oriented language for end users — not developer changelogs. Do NOT use for raw CHANGELOG.md entries (use changelog-writer) or internal dev docs.

Triages GitHub issues and pull requests: applies labels, assigns reviewers, drafts triage comments, and routes work to the correct team or milestone. Use this skill whenever a user says "triage these issues", "label my PRs", "help me route this issue", "which team should own this bug?", "tag these tickets", or "help me clean up the backlog". Also activate when someone pastes a list of issues/PRs and asks for prioritization or categorization. Do NOT use for writing code fixes, closing issues without review, or making architectural decisions about issue scope.

Plans and writes responsive CSS breakpoint strategies for layouts and components: defines breakpoint scales, writes media queries, and adapts components across screen sizes. Use this skill whenever a user says "make this responsive", "add breakpoints to this layout", "how should I handle mobile vs desktop?", "write responsive media queries", "what breakpoints should I use?", "my layout breaks on mobile", "adapt this for tablet", or "plan a mobile-first responsive strategy". Also activate when someone has a desktop layout and wants to adapt it for smaller screens. Covers plain CSS, Tailwind, and CSS-in-JS media queries. Do NOT use for full-page layout generation (use css-layout-helper) or viewport meta tags configuration.