Skills

Biologically-inspired project memory system (hippo-memory) that gives Claude persistent, decay-based memory across sessions. Use this skill whenever you need to remember something for next time, recall what you learned before, set up project memory on a new machine, initialize hippo in a project, check memory health, record architectural decisions, learn from git history, or import memories from other tools. Also trigger on: "what did I learn", "have we seen this before", "remember this", "set up memory", "init hippo", "project memory", "cross-session context", "hippo", "memory consolidation", "decay-based memory", or when starting work on a project that has a .hippo/ directory.

Read, search, send, reply, forward, and manage emails across multiple Gmail accounts via the Himalaya CLI (IMAP/SMTP). Unlike API-based approaches (Google REST API, MCP Gmail tools), this skill uses direct IMAP/SMTP protocol access — no bun runtime, no API quotas, no draft-only limitations. Supports full send capability, multi-account switching, folder management, flags, attachments, and threaded conversations. Use this skill for ALL email operations — reading mail, searching inbox, sending emails, replying to threads, forwarding messages, managing folders/labels, downloading attachments, and checking multiple accounts. Trigger on any mention of email, inbox, Gmail, sending mail, checking mail, replying, forwarding, or email search.

Web scraping, search, crawling, and page interaction via the Firecrawl CLI. Use this skill whenever the user wants to search the web, find articles, research a topic, look something up online, scrape a webpage, grab content from a URL, extract data from a website, crawl documentation, download a site, or interact with pages that need clicks or logins. Also use when they say "fetch this page", "pull the content from", "get the page at https://", or reference scraping external websites. This provides real-time web search with full page content extraction and interact capabilities — beyond what Claude can do natively with built-in tools. Do NOT trigger for local file operations, git commands, deployments, or code editing tasks.

Save critical working context to structured markdown files before an agent's context window fills up, so future agents can resume without repeating failed approaches or losing hard-won knowledge. Use this skill whenever you notice your context is getting large, when a manager agent tells you to checkpoint, when you are about to be replaced by a fresh agent, or when you have accumulated significant learnings that would be costly to rediscover. Also trigger when anyone mentions context checkpoint, save context, preserve context, session handoff, context running low, or context continuity.

Control the cmux terminal multiplexer on macOS via its CLI. Use when managing workspaces, splitting panes, sending text/keys to terminals, reading terminal output, opening and automating browsers, sending notifications, or displaying status indicators in the cmux sidebar. Requires cmux to be running.

Create, read, modify, and inspect PowerPoint (.pptx) presentations. Use when the user's task involves slide decks, presentations, or .pptx files — whether creating from scratch, editing templates, extracting content, or generating visual previews. Handles the full lifecycle including text extraction, slide manipulation, thumbnail generation, orphan cleanup, and quality checks.

Search Google Flights for flight prices, routes, and deals via SerpAPI. Use this skill whenever the user asks about flights, airfare, travel between cities, booking flights, comparing flight prices, finding cheap flights, or planning air travel. Trigger on phrases like 'find me a flight', 'how much to fly to', 'flights from X to Y', 'cheapest flight', 'nonstop to', 'round trip to', or any mention of airports, airlines, or air travel pricing. This is the only way to get real-time flight pricing data.

Design rules and visual standards for creating professional, minimal presentation slides. Use this skill whenever an agent is designing PowerPoint content — choosing layouts, typography, color palettes, text density, or image placement. This skill does NOT handle the technical mechanics of creating .pptx files (use the pptx skill for that). Instead, it provides the design constraints that produce slides experts would recognize as professionally designed. Reference this skill proactively when generating any presentation content, slide outlines, or visual layouts.

Set up and manage Oathbound — the trust and verification layer for Claude Code skills. Use this skill whenever the user mentions oathbound, skill verification, skill trust, publishing skills, installing verified skills, setting up skill security, or when initializing a new project that will use third-party Claude Code skills. Also use this skill proactively when setting up any project that uses `.claude/skills/` — every project with third-party skills should be oathbound to prevent supply-chain attacks from unverified code running inside an agent's context.

Web search with full page content extraction. Use this skill whenever the user asks to search the web, find articles, research a topic, look something up, find recent news, discover sources, or says "search for", "find me", "look up", "what are people saying about", or "find articles about". Returns real search results with optional full-page markdown — not just snippets. Provides capabilities beyond Claude's built-in WebSearch.

Extract clean markdown from any URL, including JavaScript-rendered SPAs. Use this skill whenever the user provides a URL and wants its content, says "scrape", "grab", "fetch", "pull", "get the page", "extract from this URL", or "read this webpage". Handles JS-rendered pages, multiple concurrent URLs, and returns LLM-optimized markdown. Use this instead of WebFetch for any webpage content extraction.

Discover and list all URLs on a website, with optional search filtering. Use this skill when the user wants to find a specific page on a large site, list all URLs, see the site structure, find where something is on a domain, or says "map the site", "find the URL for", "what pages are on", or "list all pages". Essential when the user knows which site but not which exact page.

Download an entire website as local files — markdown, screenshots, or multiple formats per page. Use this skill when the user wants to save a site locally, download documentation for offline use, bulk-save pages as files, or says "download the site", "save as local files", "offline copy", "download all the docs", or "save for reference". Combines site mapping and scraping into organized local directories.

Bulk extract content from an entire website or site section. Use this skill when the user wants to crawl a site, extract all pages from a docs section, bulk-scrape multiple pages following links, or says "crawl", "get all the pages", "extract everything under /docs", "bulk extract", or needs content from many pages on the same site. Handles depth limits, path filtering, and concurrent extraction.

DEPRECATED — use scrape + interact instead. Interact lets you scrape a page and then click, fill forms, and navigate without managing sessions manually. Use this skill when the user needs to interact with a webpage, log into a site, click buttons, fill forms, navigate multi-step flows, handle pagination, or when regular scraping fails because content requires JavaScript interaction. Triggers on "click", "fill out the form", "log in to", "paginated", "infinite scroll", "interact with the page", or "scrape failed".

AI-powered autonomous data extraction that navigates complex sites and returns structured JSON. Use this skill when the user wants structured data from websites, needs to extract pricing tiers, product listings, directory entries, or any data as JSON with a schema. Triggers on "extract structured data", "get all the products", "pull pricing info", "extract as JSON", or when the user provides a JSON schema for website data. More powerful than simple scraping for multi-page structured extraction.

Query, create, and manage Linear issues, projects, teams, and workflows via the Linear GraphQL API. Use this skill for ALL Linear operations — searching issues, viewing issue details, creating and updating issues, adding comments, listing teams, projects, workflow states, labels, and cycles. Trigger on any mention of Linear, issue tracking, sprint planning, or project management via Linear.

Create and manage virtual debit cards for making online purchases using the agent-cards CLI. ALWAYS use this skill when the user asks you to buy something online, pay for a service, purchase API credits, register a domain, subscribe to a SaaS tool, make an x402 payment, or any task that requires spending money. Also use when the user mentions agent-cards, virtual cards, prepaid cards, spending limits, or payment credentials. This skill covers the full lifecycle: creating funded cards, checking balances, retrieving payment credentials (PAN/CVV/expiry), filling payment forms via browser automation, making x402 HTTP payments, and closing cards when done.

Remove backgrounds from images using AI-powered rembg models (photos, portraits, products) and threshold-based removal (line art, drawings, logos). ALWAYS use this skill when the user wants to remove a background, make an image transparent, cut out or isolate a subject, process product photos for e-commerce, remove backgrounds from headshots or portraits, extract foreground elements, or batch-process a folder of images for background removal. Also use when someone mentions transparent PNG, alpha channel extraction, or wants a drawing/sketch on a transparent background instead of white. This skill provides a Python script with model selection, alpha matting for fine edges (hair, fur), and batch processing — do NOT attempt background removal manually without it.

Control and automate Chrome browser sessions using Puppeteer. ALWAYS use this skill when users need browser automation, web scraping, page inspection, screenshot capture, form filling, accessibility auditing, network monitoring, performance tracing, or any interaction with a live web page. Covers the full scope of Chrome DevTools MCP: navigation, input automation (click, fill, type, drag, upload), emulation (viewport, network throttling, geolocation, dark mode), debugging (screenshots, a11y snapshots, console logs, JS evaluation), network inspection, performance tracing, and memory snapshots. Trigger on any mention of: browser automation, puppeteer, click a button, fill a form, take a screenshot of a website, check console errors, inspect a page, test a webpage, scrape a site, emulate mobile, check accessibility, audit a page, capture network requests, or performance profiling.

Create interactive 3D scenes, visualizations, and experiences using Three.js. ALWAYS use this skill when users ask for 3D graphics, Three.js scenes, WebGL content, 3D product viewers, 3D data visualizations, 3D animations, or any browser-based 3D rendering. Also use when users mention orbit controls, GLTF models, raycasting, PBR materials, or scene composition — even if they don't say "Three.js" explicitly.

Analyze Docker configurations for anti-patterns, debug container networking, optimize Compose setups, inspect logs, and troubleshoot container lifecycle. ALWAYS use when users mention Dockerfile optimization, container networking issues, Docker Compose problems, port conflicts, volume mounting errors, container exit codes, or health check setup. This is NOT for Kubernetes — it focuses exclusively on Docker and Docker Compose diagnostics.

Process, resize, convert, and optimize images. ALWAYS use when users need to resize images, convert between formats (PNG, JPG, WebP, AVIF), optimize file sizes, generate favicons, or inspect image metadata. Use this instead of manual ImageMagick or Sharp commands.

Debug and validate GitHub Actions workflows — catch trigger condition mistakes, validate matrix builds, optimize caching, audit secrets and permissions. ALWAYS use when workflows dont trigger, matrix builds fail, caching isnt working, scheduled jobs dont run, or users need workflow debugging, CI/CD optimization, or Actions security review.

Manage Supabase projects — execute SQL queries, push migrations, generate TypeScript types, search Supabase docs, view project logs, manage edge functions, and run security/performance checks. ALWAYS use this skill whenever the user mentions Supabase, Supabase databases, Supabase edge functions, Supabase migrations, Supabase RLS, Supabase logs, or any Supabase project management. Also use when querying a Supabase-hosted database, deploying to Supabase, checking Supabase security advisors, generating types from a Supabase schema, or searching Supabase documentation — even if the user doesn't explicitly say "use the supabase skill".

Push the current branch and create a pull request with a conventional commit title, summary, and optional test plan. Use this skill for ALL pull request operations — creating PRs, pushing branches for review, opening merge requests, submitting code for review, or any workflow that ends with a PR on GitHub. Trigger on any mention of creating a PR, opening a pull request, pushing for review, submitting changes, "make a PR", "open a PR", "push and create PR", "send this for review", or preparing a branch for merge. Also trigger when the user says "PR this", "ship it", "send it up", or asks to get their changes reviewed.

Send and receive messages via the Telegram Bot API. Use this skill for ALL Telegram operations — sending messages, reading recent messages/updates, sending photos or documents, forwarding messages, replying to messages, and checking bot info. Trigger on any mention of Telegram, sending Telegram messages, bot messages, chat notifications, Telegram bot, notifying someone via Telegram, or wanting to push a message to a chat. Also use when the user mentions a Telegram chat ID, wants to check for incoming Telegram messages, or needs to send files/images through Telegram.

Web UI design philosophy — layout, typography, color, responsive design, visual hierarchy, whitespace, accessibility, and design system coherence. ALWAYS consult before writing CSS, choosing spacing, typography scales, colors, breakpoints, or layouts. Contains specific systems (8px grid, modular type scales, 60-30-10 color rule, WCAG contrast, mobile-first breakpoints) that produce professional results. Trigger on CSS styling, layout, spacing, typography, fonts, colors, dark mode, responsive design, mobile layout, accessibility, visual hierarchy, whitespace, design system, component styling, UI review, or web page aesthetics. This skill has opinionated constraints that override generic design knowledge.

ALWAYS use this skill when the user asks you to check, inspect, or verify a running web page or dev server. This skill provides a CLI script (bun .claude/skills/visual-feedback/script.ts) that connects to a local Vite dev server to capture accessibility trees, console errors, failed network requests, and screenshots. Trigger on any mention of: checking a page, inspecting the DOM, verifying UI changes, taking a screenshot of a route, checking for console errors on a page, responsive testing across viewports, checking page structure, a11y inspection, or viewing what a page looks like. Also trigger when the user says 'check the page', 'look at the page', 'what does it look like', 'screenshot this route', 'inspect the layout', 'any errors on the page', or asks to verify changes they just made to the UI. This is NOT for writing Playwright/Cypress tests, scraping external sites, or reviewing source code — it specifically inspects the live running dev server.

Use this skill whenever writing, editing, reviewing, or debugging Svelte or SvelteKit code. This skill contains critical project-specific conventions and Svelte 5 runes patterns that you MUST consult — without it you will produce broken Svelte 4 code (export let, on:click, slot, $:, createEventDispatcher) instead of correct Svelte 5 code ($props, $state, $derived, $effect, $bindable, {@render}, snippets). Also covers shadcn-svelte component patterns, tailwind-variants tv() usage, and Tailwind CSS 4.x configuration. Trigger on any mention of: Svelte, SvelteKit, .svelte files, runes, $state, $derived, $effect, $props, $bindable, snippets, {@render}, shadcn-svelte, svelte-check, +page.svelte, +layout.svelte, +server.ts, load functions, form actions, or migrating from Svelte 4 to Svelte 5.

ALWAYS use this skill when users mention OpenSCAD, .scad files, 3D modeling code, CAD primitives, or 3D printing geometry. This skill provides a validated code generation pipeline via bun script.ts that checks parameter types, generates correct syntax, and can compile output with OpenSCAD to verify correctness — do NOT write OpenSCAD code from memory, as subtle parameter bugs (wrong named args, missing semicolons, incorrect extrusion syntax) are common and hard to catch without compilation. Covers all 20 primitives (cube, sphere, cylinder, circle, square, polygon, text, union, difference, intersection, translate, rotate, scale, mirror, color, linear_extrude, rotate_extrude, hull, minkowski, offset). Use this skill even for simple OpenSCAD questions — the info action provides exact signatures and parameter details that prevent mistakes.

Interact with Notion pages, databases, and content via the Notion REST API. Use this skill for ALL Notion operations — searching pages, reading content, listing and querying databases, creating and updating pages. Trigger on any mention of Notion, workspace, pages, databases, or content management via Notion.

Fetch up-to-date documentation and code examples for any programming library using the Context7 API. Use this skill whenever the user asks to look up docs, fetch API references, find code examples, or get current documentation for any npm/programming library. Trigger on phrases like 'look up the docs for X', 'fetch documentation', 'what's the API for X', 'get current docs', or any request to retrieve library documentation.

Interact with Anki flashcards via the AnkiConnect REST API. Use this skill for ALL Anki flashcard operations — creating cards, searching decks, checking review stats, syncing with AnkiWeb, listing note types, and updating existing notes. Trigger on any mention of Anki, flashcards, spaced repetition, AnkiConnect, decks, or card review. Also use when the user wants to turn content into flashcards, even if they don't explicitly mention Anki.

Query, analyze, and manage PostHog analytics data — run HogQL queries, browse events, look up users, search session recordings, manage feature flags, and retrieve insights/dashboards. Use when users mention PostHog, analytics queries, event tracking, feature flags, session recordings, or HogQL.

Brand design system and visual standards for Oathbound. Use this skill whenever creating any visual output -- slides, documents, web pages, HTML artifacts, posters, or diagrams -- that should follow Oathbound brand identity. Provides exact color values, typography specs, spacing rules, and component patterns extracted from the Oathbound codebase. Use this skill proactively whenever the user mentions Oathbound, oathbound.ai, skill attestation, or asks for branded output for Oathbound, even if they don't say "brand guidelines."

Brand design system and visual standards for Cantaloupe AI. Use this skill whenever creating any visual output — slides, documents, web pages, HTML artifacts, posters, or diagrams — that should follow Cantaloupe AI brand identity. Provides exact color values, typography specs, spacing rules, and component patterns extracted from the Cantaloupe codebase and marketing site. Use this skill proactively whenever the user mentions Cantaloupe, trycantaloupe.com, Union, or asks for branded output for their company, even if they don't say "brand guidelines."

Read, create, edit, format, and validate Excel workbooks (.xlsx, .xlsm). Ingest CSV/TSV files and convert to spreadsheets. Recalculate formulas via LibreOffice and scan for errors. Use when the user supplies or requests a spreadsheet file, or when building/updating/examining tabular data destined for Excel output.

Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs.

Suite of tools for creating elaborate, multi-component claude.ai HTML artifacts using modern frontend web technologies (React, Tailwind CSS, shadcn/ui). Use for complex artifacts requiring state management, routing, or shadcn/ui components - not for simple single-file HTML/JSX artifacts.

Knowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack."

Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.

Create, edit, validate, and convert Word documents (.docx). Use when building documents programmatically, extracting content from existing documents, making tracked-change edits to document XML, adding comments, converting between .doc/.docx/PDF formats, or validating OOXML structure. Supports full unpack-edit-pack workflows with differential validation.

Build apps with the Claude API or Anthropic SDK. TRIGGER when: code imports `anthropic`/`@anthropic-ai/sdk`/`claude_agent_sdk`, or user asks to use Claude API, Anthropic SDKs, or Agent SDK. DO NOT TRIGGER when: code imports `openai`/other AI SDK, general programming, or ML/data-science tasks.

Creating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems. Create original algorithmic art rather than copying existing artists' work to avoid copyright violations.

Handle all PDF operations - reading, extracting, creating, filling forms, securing, and converting PDF documents. Use when the user mentions .pdf files, asks to extract text/tables/images, fill forms, merge/split pages, add watermarks, encrypt/decrypt, or convert PDFs to images.

Generates a formatted changelog from recent git commits. Groups commits by type (feat, fix, chore, docs, refactor) and outputs clean markdown. Use when you need to summarize recent changes for a release, PR description, or team update.