Skills
99,064 packaged capabilities and playbooks for agents. Install any of them with one command.
Development
Best practices for using vectors and AI with MariaDB. Use when writing SQL involving VECTOR columns, building RAG or semantic search with MariaDB, choosing embedding models, designing vector indexes, or integrating MariaDB with AI frameworks like LangChain, LlamaIndex, or Spring AI. Also use when the user mentions MariaDB and any of: vectors, embeddings, similarity search, nearest neighbor, AI, RAG, or LLM. IMPORTANT — MariaDB has native built-in vector support since version 11.7; it is NOT a plugin or extension (unlike pgvector for PostgreSQL). Do not recommend installing any extension.
Midnight on-chain execution — WASM runtime, FRAME pallets, pallet-midnight ledger state machine, proof-based transaction validation, and Patricia-Merkle state commitments. Use when explaining runtime architecture, pallet roles, or how contract/ZSwap txs update chain state.
Train machine learning models end-to-end, covering data loading, preprocessing, architecture selection, training loops, validation, and checkpointing. Use when the user requests model training or provides relevant inputs for this workflow.
Design and implement Monte Carlo methods for uncertainty quantification, risk analysis, and probabilistic simulations across scientific and financial domains. Use when "monte carlo, random sampling, uncertainty quantification, risk analysis, stochastic simulation, MCMC, variance reduction, probabilistic, " mentioned.
MyBatis-Plus enhanced ORM skill. Covers choosing between LambdaQueryWrapper and QueryWrapper (always Lambda), pagination plugin (PaginationInnerInterceptor) configuration, optimistic locking (@Version + OptimisticLockerInnerInterceptor), logical deletion (@TableLogic global/local configuration), auto-fill (@TableField fill + MetaObjectHandler), choosing between ActiveRecord and Mapper modes, preventing full-table updates (BlockAttackInnerInterceptor), and extracting a generic PageQuery. Use when the user performs database operations with MyBatis-Plus, configures pagination or optimistic locking, or chooses a query approach. Complements the mybatis skill: mybatis focuses on XML/ResultMap/SQL style, while mybatis-plus focuses on enhanced features.
Execute a comprehensive NestJS Project Health Audit. Analyzes tech stack, architecture, API design, data layer, testing, code quality, CI/CD, and documentation. Produces a shareable Markdown report with section scores and weighted overall score. Use when the user asks to audit a NestJS project, run a health check, evaluate backend quality, or assess technical debt. Triggers on: 'nestjs audit', 'health audit', 'backend audit', 'nestjs health', 'node audit', 'api audit', 'project quality check'.
Use when Obsidian note automation runs in cron/headless environments and obsidian-cli emits URI failure signatures (for example, `Failed to execute Obsidian URI`) that may not set a non-zero exit code. Detect false-success cases, fallback to deterministic markdown file writes, and record traceable fallback paths in run artifacts.
Search relevance and ranking on OpenSearch/Elasticsearch for a two-sided marketplace — candidate retrieval (hybrid BM25 + kNN, RRF, two-tower EBR), base relevance (BM25F, multi_match, LambdaMART), quality signals (Wilson lower bound, Bayesian average, rank_feature saturation/sigmoid), personalization (listing/user/session embeddings), spatial/temporal decay (gauss/exp), marketplace balance (conversion-weighted ranking, supply fairness, Pareto multi-objective), bias correction (IPS, click models, Thompson sampling), empirical evaluation (judgment sets, NDCG, ablation, A/B sizing, CUPED, regression suites), and diversity (MMR, DPP, max-per-host). Triggers on function_score, rank_feature, script_score, kNN, hybrid query, learning-to-rank, two-sided ranking, exposure fairness, NDCG, A/B testing, judgment set construction, ranking ablation, or "why is my OpenSearch ranking bad". Applies to Elasticsearch too — same APIs.
Run, verify, reseed, and repair Paperclip isolated dev workspace services. Use when asked to start or fix a managed project/worktree service and prove health, login readiness, cloned data, runtime visibility, and correct port ownership.
Finish a 3D scene's look with the postprocessing library — a tasteful pass (ambient occlusion, subtle bloom, tone mapping, color grading) plus the correct WebGLRenderer attributes. Use to make a scene read as a polished game rather than a raw render, or for any screen-space image effect.
React and Next.js performance optimization guidelines from Vercel Engineering. This skill should be used when writing, reviewing, or refactoring React/Next.js code to ensure optimal performance patterns. Triggers on tasks involving React components, Next.js pages, data fetching, bundle optimization, or performance improvements.
Debug and fix a Sentry issue — find it (by link, ID, or search), pull full context (stack trace, breadcrumbs, trace, logs), optionally run Seer root-cause / autofix, apply the code fix, and resolve it via a `Fixes PROJECT-NAME-12A` commit/PR. Use when working a known error or hunting one down to fix.
Set up Sentry releases and deploy tracking — tag events with a version and environment, create the release in CI with its commits, and wire up suspect commits and code mappings, so Sentry can show which release introduced an issue, which commit is responsible, and release health. Use when asked to set up releases, track deploys, see what changed, or when issues show an unknown release or no suspect commit.
[AUTO-INVOKE] MUST be invoked when debugging failed on-chain transactions. Covers transaction receipt analysis, gas diagnosis, calldata decoding, revert reason extraction, and state verification. Trigger: any task involving failed tx analysis, revert debugging, or on-chain transaction troubleshooting.
Create, validate, and transition documentation artifacts (Vision, Initiative, Epic, Spec, Spike, ADR, Persona, Runbook, Design, Journey) through lifecycle phases. Handles spec writing, feature planning, epic creation, initiative creation, ADR drafting, research spikes, persona definition, runbook creation, design capture, architecture docs, phase transitions, implementation planning, cross-reference validation, and audits. Also invoke to update frontmatter fields, re-parent an artifact under a different epic or initiative, or set priority on a Vision or Initiative. Chains into swain-do for implementation tracking on SPEC; decomposes EPIC/VISION/INITIATIVE/JOURNEY into children first.
Headless UI for virtualizing scrollable elements in Vue. Use when writing, reviewing, or debugging code that imports @tanstack/vue-virtual, calls useVirtualizer or useWindowVirtualizer, or virtualizes lists, grids, masonry, window-scrolled content, or chat feeds in Vue. Provides version-accurate adapter rules, Virtualizer options and methods, and rendering recipes for @tanstack/vue-virtual 3.13.x on @tanstack/virtual-core 3.17.x.
Python type checking expertise using ty - the extremely fast type checker by Astral. Use when: (1) Adding type annotations to Python code, (2) Fixing type errors reported by ty, (3) Migrating from mypy/pyright to ty, (4) Configuring ty for projects, (5) Understanding advanced type patterns (generics, protocols, intersection types), (6) Setting up ty in editors (VS Code, Cursor, Neovim, PyCharm).
JS/TS best practices and rule enforcement — JS idioms, TS type safety, Zod runtime validation, error handling, and Deno workflows. Trigger when working with TS/JS code; discriminated unions, generics, narrowing, any vs unknown; Zod parse/safeParse; Result types; readonly, `as const`, modules; destructuring, optional chaining; pipe/compose/currying; unit tests, mock vs stub; Deno (deno.json, permissions, JSR, Deno.serve, Node-to-Deno migration); or "is this idiomatic JS/TS". Do NOT use for advanced generic/conditional/mapped types or compiler-error puzzles (`typescript-magician`), class/SOLID design (`object-oriented-programming`), or language-agnostic test strategy (`testing`)
Select, integrate, and operate multi-model video-generation gateways — hosted inference aggregators such as fal.ai, Replicate, and WaveSpeed that expose many third-party video (and video-adjacent audio) models behind one account, one API surface, and one bill. Use when deciding whether to route video generation through a gateway versus a direct model-provider API; when comparing gateways on catalog, async/queue job design, webhooks, pricing, schema discovery, and version pinning; and when operating gateway workloads in production — spend controls, retries, cross-gateway failover, content-safety differences, data retention, and commercial-rights passthrough of the underlying model. Not for direct model-provider APIs, local/self-hosted inference, model training/fine-tuning, or image-only gateway use.
Writing webpack 5 plugins — hook selection (compiler vs compilation, tap vs tapAsync, processAssets stages), the asset pipeline (emitAsset, source classes, info metadata, source maps), watch-mode and persistent caching (file/context/missing/buildDependencies), plugin lifecycle (constructor purity, multi-compiler isolation, shutdown cleanup), schema-utils validation, WebpackError reporting, jest-worker parallelism, and compatibility patterns (compiler.webpack namespace, peerDependencies, getCompilationHooks WeakMap). Patterns are drawn from production plugins like mini-css-extract-plugin, terser-webpack-plugin, compression-webpack-plugin, and Next.js's webpack plugins. Trigger when writing, reviewing, or debugging webpack 5 plugins — even if the user doesn't explicitly mention "best practices" — anytime an `apply(compiler)` method is being written, hooks are being tapped, or a plugin imports from `webpack-sources`, the rules in this skill apply.
Use whenever you face a webpack-build problem that ends with needing to write a plugin — 26 production-shaped recipes covering bundle-size budgets, forbidden architectural imports, env-var validation, secret-leak prevention, build-info injection, asset manifests, license walking, SRI hashes, virtual modules (Vite-style), filesystem routing (Next.js-style), generated barrels, runtime-driven TS types, library replacement (react→preact), debug-stripping, conditional polyfills, dynamic banners, feature flags, build-duration regression tracking, desktop/Slack notifications, changed-chunks diffing, browser auto-open, gzip/brotli pre-compression, image optimization, type-based dist layout, empty-chunk cleanup, cache-busting query strings. Each recipe is a complete working plugin drawn from production patterns at Next.js, Storybook, webpack-contrib. Trigger when the user asks how to do X with webpack or whether a plugin exists for X. Companion to webpack-plugin-authoring.
Modernizes Angular code such as components and directives to follow best practices using both automatic CLI migrations and Bitwarden-specific patterns. YOU must use this skill when someone requests modernizing Angular code. DO NOT invoke for general Angular discussions unrelated to modernization.
Use when editing .astro/.mdx files, astro.config.*, Astro content collections, client directives, Actions/forms, Tailwind CSS, view transitions, server features, adapters, or upgrades to Astro 7. Routes to current Astro 7 and Zod 4 patterns while guarding against removed Astro 3/4/5/6 APIs. NOT for generic framework-agnostic frontend design or non-Astro projects.
Use when authoring (or amending) a technical-design document (a TDD / engineering design doc / design RFC) for one feature or component — the detailed implementation design for building it within an existing system. Guides the METHOD, not the outline: grounding the design in established practice and the project's real constraints, tracing every decision bidirectionally to a requirement, comparing at least one real (non-strawman) alternative with a stated decision criterion, referencing the architecture-doc / API spec / data-model rather than duplicating them, naming the failure modes, the observability signals, the testing, and the rollout — and amending an approved design as a versioned, ripple-analyzed delta — to a bar where an engineer can implement without re-deriving the design. Assumes the approved PRD + feature-spec as upstream input, never a blank page. Not for system-wide architecture, not for the API contract or data schema, and not for reviewing a finished TDD.
Autonomous PR monitor — polls every 2 minutes for merge conflicts, CI/CD failures across GitHub Actions, Buildkite, Vercel, and Fly.io, review comments, and merge readiness. Auto-detects PR from current branch, fixes what it can, notifies on state changes. No setup questions. Use when asked to babysit a PR, watch a PR, monitor CI, keep a PR green, handle merge conflicts, or poll PR status
Implement Beads or beads_rust work items from an issue id. Use when the user runs $beads-implement <bead-id> or asks to implement a bead/epic; if it is an epic, implement all sub-tasks, committing after each one and only interrupting when blocked or a decision is needed.
Designs new software features and multi-component changes for the current repo into implementation-ready BDD specs via autonomous codebase research, then commits the design for review (runs to completion without mid-design questions). Scope covers new features, greenfield design, and multi-component changes. Not for hardware/physical-system design, out-of-repo questions, single-file refactors, or known-root-cause bugs (those go to direct edits or systematic-debugging).
Cameras for Three.js games — viverse's ready-made character camera (third/first-person orbit, collision, zoom) and, when you need it, building cameras and effects (screen shake, FOV speed) from scratch as ECS systems. Use when implementing or tuning camera movement and behavior.
Use when writing new code, naming modules/files/functions/variables, reviewing or refactoring for readability, reducing complexity, deciding whether an abstraction belongs, or reviewing with CUPID; also the constructive hacker ethos (curiosity, craftsmanship, open-source contribution, remove drudgery; never unauthorized access). Behavior-preserving refactors within a function, class, or file route to the `refactoring/` sub-skill. Pairs with language skills (typescript, go, lisper). DO NOT USE for behavior changes (`testing` (TCRDD reference)), cross-file dependency untangling (`mikado-method`), or pure simplicity/YAGNI debates (`simple`)
Synchronize Comet's Chinese and English docs, README, Skills, release notes, and website pages while preserving confirmed semantics, repository structure, submodule boundaries, and focused validation. Use when documentation changes must stay bilingual, or when Chinese wording needs a corresponding English version.
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.