Skills
98,873 packaged capabilities and playbooks for agents. Install any of them with one command.
All
Ultra-compressed commit message generator. Cuts noise from commit messages while preserving intent and reasoning. Conventional Commits format. Subject ≤50 chars, body only when "why" isn't obvious. Use when user says "write a commit", "commit message", "generate commit", "/commit", or invokes /caveman-commit. Auto-triggers when staging changes.
Compress natural language memory files (CLAUDE.md, todos, preferences) into caveman format to save input tokens. Preserves all technical substance, code, URLs, and structure. Compressed version overwrites the original file. Human-readable backup saved as FILE.original.md. Trigger: /caveman-compress FILEPATH or "compress memory file"
This skill provides character arc development frameworks and techniques. Covers the core arc formula (flaw to strength, want to need, lie to truth), positive change arcs, negative arcs, and supporting character development. Use when: developing character transformations, tracking character growth, analyzing protagonist journeys, or building supporting character arcs.
Use only when working on the docs site. Put a chart of real numbers on a docs page with the `<Chart />` block, a bar, line, area, pie or radar chart in the brand's palette. Use whenever a page compares, ranks, breaks down or tracks numbers, or when a paragraph carries three sentences of figures.
Skill dedicated to a tile and wood flooring buying guide. Covers wood-grain tiles and continuous laying, installation techniques, wood flooring selection, skirting boards and threshold stones, kitchen and bathroom tiles, and tile selection across the full process, integrating the highlights of the Xi'an Renovation Classroom member knowledge base in reverse order to help users buy tiles and wood flooring with less hassle and fewer pitfalls.
Creates and reviews CLAUDE.md configuration files for Claude Code. Applies HumanLayer guidelines including instruction budgets (~50 user-level, ~100 project-level), WHAT/WHY/HOW framework, and progressive disclosure. Identifies anti-patterns like using Claude as a linter for style rules.
Corrective cleanup of AI-generated code — removes LLM-specific patterns while preserving behavior. Use when the user says "deslop", "slop", "clean AI code", "remove AI patterns", or when you spot LLM-generated code smells after any generation session. For general post-change review and fixes ("clean up", "simplify"), use the simplify skill instead.
OpenAI's model connecting vision and language. Enables zero-shot image classification, image-text matching, and cross-modal retrieval. Trained on 400M image-text pairs. Use for image search, content moderation, or vision-language tasks without fine-tuning. Best for general-purpose image understanding.
Evaluate Clojure skills through rigorous real-project testing with and without skills enabled. Use this skill whenever you need to assess whether a Clojure skill (like clj-discover, clj-debug, clj-replace) actually improves development outcomes. Compare behavior differences via context-logs, code quality, documentation depth, and test comprehensiveness. This prevents vague or hypothetical evaluation and forces concrete real-world testing.
Spin up an isolated cloud sandbox — a full root Ubuntu VM (4 vCPU / 8 GB / 80 GB), not a thin isolate or container — to run code, builds, installs, or shell commands off the user's machine. Billed per second, ~10x cheaper than E2B/Modal; Docker and every major toolchain preinstalled. Use whenever work should not touch the host filesystem.
Use when a CloudCannon site uses Bookshop — `*.bookshop.yml` files, `@bookshop/*` packages, `_bookshop_name` keys, `bookshop:live`, or `partial "bookshop"` calls or `{% bookshop %}` tags. Covers how Bookshop works, maintaining a site that stays on it, and migrating it to editable regions.
Connect Workers to PostgreSQL/MySQL with Hyperdrive's global pooling and caching. Use when: connecting to existing databases, setting up connection pools, using node-postgres/mysql2, integrating Drizzle/Prisma, or troubleshooting pool acquisition failures, TLS errors, nodejs_compat missing, or eval() disallowed.
Store and transform images with Cloudflare Images API and transformations. Use when: uploading images, implementing direct creator uploads, creating variants, generating signed URLs, optimizing formats (WebP/AVIF), transforming via Workers, or debugging CORS, multipart, or error codes 9401-9413.
Build Model Context Protocol (MCP) servers on Cloudflare Workers - the only platform with official remote MCP support. TypeScript-based with OAuth, Durable Objects, and WebSocket hibernation. Use when: deploying remote MCP servers, implementing OAuth (GitHub/Google), using dual transports (SSE/HTTP), or troubleshooting URL path mismatches, McpAgent exports, OAuth redirects, CORS issues.
CloudNative-PG (CNPG) PostgreSQL database management for the Kubernetes homelab. Covers shared platform cluster, dedicated per-app clusters, credential provisioning, cross-namespace replication via kubernetes-replicator, and monitoring. Use when: (1) Adding a new database for an application, (2) Creating a dedicated CNPG cluster, (3) Setting up database credentials and cross-namespace replication, (4) Debugging database connectivity or CNPG cluster health, (5) Adding PostgreSQL extensions for specialized workloads. Triggers: "database", "postgresql", "postgres", "cnpg", "cloudnative-pg", "pooler", "pgbouncer", "database credentials", "db password", "managed roles", "Database CRD", "database cluster", "shared database", "dedicated database", "cnpg cluster"
Provides comprehensive code review covering 6 focused aspects - architecture & design, code quality, security & dependencies, performance & scalability, testing coverage, and documentation & API design. Use this skill for deep analysis with actionable feedback after significant code changes.
(NS) Senior Tech Lead code review on SOLID, clean code, performance, security, and testability. Use proactively after writing or modifying code, before opening PRs, after implementation closure, or when the user asks for a code review, PR review, or issue review gate — even if they do not name this skill. For GitLab issue execution with ISSUE_URL, use Issue review mode. Do NOT use for root-cause debugging (use code-investigator).
Use this when user wants you to walk through (code or text) files in a EDITOR to either explain how some code works, or to show the user what changes you made, etc. You would typically use this repeatedly to show the user your changes or code files one by one, sometimes with specific line-numbers. This way the user is easily able to follow along in their favorite EDITOR as you point at various files possibly at specific line numbers within those files.
Comment-hygiene-only reviewer — flags ephemeral review-ID references, historical change-narration, stale comments, reviewer-appeasement, redundant restating, verbose rationale essays, and diff-level comment density in the scoped diff, and normalizes findings into the verify pipeline format
Teaches AI assistants how to develop FlutterFlow apps using MCP tools. Use this skill when working with FlutterFlow projects, editing FF YAML, creating or inspecting pages and components, reading project configuration, or navigating FlutterFlow widget trees. It covers all 25 MCP tools for discovery, reading, editing, and settings. Triggers on: FlutterFlow, FF YAML, FF page, FF component, FF widget, FF theme, FF project.
Expert in running the day-to-day operations of thriving communities. Covers moderation systems, onboarding flows, crisis management, scaling operations, and team management. The engine room that keeps communities healthy and safe. Use when "community operations, community moderation, community management, onboarding flow, crisis management, trust and safety, community team, " mentioned.
Generates business/company names across 10 categories (Descriptive, Metaphoric, Invented, Founder-based, Acronym, Compound, Foreign, Playful, Geographic, Legacy) with USPTO trademark screening, domain availability checking, and 0-100 scoring. Use when users need company/product/brand naming for new business launches, rebranding, trademark strategy, IP protection naming, evaluating current name strength, or any naming/branding tasks requiring systematic analysis with legal clearance.
Deep BD-grade research report on any target company — financials, leadership, hiring, tech stack, competitors, outsourcing exposure, KServe service fit, ICP score, and BD briefing. Triggers on any BD context where a company name is dropped — even without an explicit "research" request. "Look up [company]", "check out [company]", "who should I call at [company]", "what do we know about [company]", "get me a profile on [company]", or simply naming a company while discussing a prospect, outreach plan, or account intelligence. If the context is prospecting, pre-call prep, or finding outsourcing clients — trigger immediately.
Find leads by scraping engagers from a competitor's top LinkedIn posts. Given one or more company page URLs, scrapes recent posts, ranks by engagement, selects the top N, extracts all reactors and commenters, ICP-classifies, and exports CSV. Use when someone wants to "find leads engaging with competitor content" or "scrape people who interact with [company]'s LinkedIn posts".
Use this skill to set up Salesforce Consumer Goods (CG) Cloud Mobile Sync end-to-end for the Sync Management App managed package — it runs all three acts in order (Act 1 Setup Sync, Act 2 Assign Users, Act 3 Plan & Verify), not just installing the baseline, so a sales rep can see the day's planned visits offline. Use it when asked to \"set up CG Cloud mobile sync\", \"configure Consumer Goods mobile sync\", \"set up sync for Retail Execution\", \"install the sync baseline\", \"assign a user to the sync configuration\", \"verify the sync setup\", or \"why can't my rep see their planned visits offline\" when sync setup or the baseline install is suspected. DO NOT TRIGGER for general CG Cloud mobile-app troubleshooting unrelated to sync setup, for authoring new or custom sync baselines (this installs only shipped baselines discovered at runtime), or for building or customizing the mobile app UI itself — those are out of scope.
Break saved articles, Clippings, and long foreign-language pieces into reusable Chinese material fragments, and store them reliably following an existing material library template. Whenever the user mentions "mining gold", "breaking down material", "store into the material library", "break this article into titles/quotes/structure", "batch process Clippings", or "turn English articles into Chinese material assets", prefer this skill instead of doing a plain summary.
Co-Write mode for Chinese web fiction. Use when the user wants to write together with AI instead of handing over a whole chapter; is stuck on what to write next; wants AI to continue from their own writing; wants several versions in different directions (recommended + alternatives) to choose from; or wants to set only the goal, conflict, and emotion and have AI write the prose. Three modes: Auto (AI writes automatically), Assist (I write some, AI fills in some), and Director (I set the direction, AI produces multiple versions).
SEO website crawler built on Crawl4AI. Full page audit: title, meta, H1, canonical, breadcrumbs, navigation, internal links. Site inventory, navigation audit, landing page comparison, competitor analysis. Works for Google and Yandex SEO (Cyrillic URLs, commercial factors, regionality). Integrates with yandex-search-api, yandex-metrika, yandex-webmaster, scrapedo-web-scraper. Triggers: crawl4ai, seo crawl, site audit, page inventory, site inventory, on-page audit, internal links, internal linking audit, navigation audit, landing comparison, competitor analysis, competitor pages, orphan pages, technical seo, аудит сайта, краулер, перелинковка, навигационный аудит.
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.