Skills
99,448 packaged capabilities and playbooks for agents. Install any of them with one command.
Development
One-click launch and configuration of Verl distributed training services. Trigger scenarios: (1) the user wants to start a Verl training job or deploy an RLHF/DAPO training environment (2) launching Verl training containers on an NPU cluster (3) configuring the Ray cluster and SwanLab monitoring (4) flexibly configuring acceleration features according to a 7-bit binary mask. Supports the full DAPO training workflow for Megatron models such as Qwen3-8B.
Comprehensive Vue 3 frontend development skill. Use when developing Vue 3 applications, components, or composables. Covers: (1) Creating new Vue 3 projects and components, (2) Vue 2 to Vue 3 migration, (3) Code review and optimization, (4) Composition API implementation, (5) Best practices and common patterns, (6) TypeScript integration, (7) State management with Pinia, (8) Performance optimization. Includes component templates, composable examples, migration guides, and best practices documentation.
Use to diagnose and fix incorrect gradients in differentiable Warp programs. Anything trained, optimized, calibrated, or fit through Warp kernels depends on wp.Tape gradients, so treat any misbehavior of such a workflow as a gradient problem until proven otherwise — use this when training diverges or NaNs, won't train at all, stalls or plateaus above the expected loss, converges to a wrong or biased answer, is worse than a reference implementation, works at small scale but fails at production scale, or fails a QA/validation recheck. Also for explicit symptoms — exploding, NaN/inf, zero, or subtly wrong gradients, suspected wp.Tape/backward issues, gradcheck failures — but users usually describe only the surface symptom ("the sim explodes", "the fit gets dragged toward outliers") without mentioning gradients: make that leap. Not for forward-only Warp work, build/install problems, or autograd issues in other frameworks without Warp.
Integration guide for the wujie micro-frontend framework. Covers main/sub application integration, Vue2/Vue3/React component wrapping, the startApp/setupApp/preloadApp APIs, route synchronization and navigation, keep-alive preloading, bus communication, plugins and lifecycle, sub-application refactoring, deployment troubleshooting, and common issues. Use when users mention wujie, micro-frontend, sub-application integration, or a main application loading a sub-application.
Provides guidance on coordinating multiple specialized teammates working in parallel. This skill should be used when the user needs to execute complex implementation plans, resolve cross-cutting concerns, or coordinate independent work streams requiring communication.
Code quality guidelines and best practices for writing clean, maintainable code. Covers naming, structure, error handling, testing, and code review standards. Use when writing code, reviewing, refactoring, or asking "how should I name this", "best practice for", "clean code".
Use this when user provides a Yii2 codebase for analysis or migration planning. Provides 6-phase workflow: scan, capability extraction, dependency analysis, requirements extraction, semantic indexing, and migration roadmap. Apply when user mentions Yii2, PHP legacy modernization, or framework migration to NestJS
Actorization converts existing software into reusable serverless applications compatible with the Apify platform. Actors are programs packaged as Docker images that accept well-defined JSON input, perform an action, and optionally produce structured JSON output.
Audit a whole PR against the delivery contract and return MERGE-READY or evidenced blockers with the full URL. Consumes the current review-change REVIEW-PASS receipt instead of re-running review axes; posts a SHA-bound ready comment; never edits or merges. Triggers: "audit-pr", "is this PR ready", "merge gate".
Implement JWT/cookie authentication and authorization in tRPC using createContext for user extraction, t.middleware with opts.next({ ctx }) for context narrowing to non-null user, protectedProcedure base pattern, client-side Authorization headers via httpBatchLink headers(), WebSocket connectionParams, and SSE auth via cookies or EventSource polyfill custom headers.
Use when wrapping UIKit views/controllers in SwiftUI, embedding SwiftUI in UIKit, or debugging UIKit-SwiftUI interop issues. Covers UIViewRepresentable, UIViewControllerRepresentable, UIHostingController, UIHostingConfiguration, coordinators, lifecycle, state binding, memory management.
Use when setting up or running linting and formatting in a JavaScript or TypeScript project with Biome — the single fast Rust tool that lints, formats, and organizes imports, replacing the ESLint + Prettier pair (the JS analog of Ruff). Covers installing Biome v2, the biome.json config (formatter, linter rule groups, the assist/import-organize actions, VCS integration, overrides, monorepo extends), the CLI (biome check --write, biome ci for CI), migrating from ESLint/Prettier, the v1->v2 deltas (e.g. --apply became --write), and authoring custom lint rules as GritQL plugins. Use when adding a lint/format gate, writing or fixing a biome.json, or porting an ESLint/Prettier setup. Not the Turborepo task wiring (compose with that skill).
Forensic bug and issue triage. Produces a graded investigation case file: symptoms, evidence (graded A/B/C by confidence), ranked hypotheses, suspected components, and a recommended planning response. INVESTIGATES and DOCUMENTS only — does NOT fix code, run tests, or implement. Hands off as a story to an external dev tool. Use when the user says: - "investigate this bug" / "triage this issue" / "debug triage" - "what's causing [symptom]" / "why is [feature] broken" - "something's wrong with [component]" / "users are reporting [problem]" - "figure out what happened" / "root cause this" - "I need a bug report" / "create an investigation case file" - "forensic analysis" / "diagnose this" - "error triage" / "issue investigation" - "what do we know about [incident/issue]". Three intents: Create (new case file), Update (add evidence to open case), Validate (check draft for completeness before handoff). Output lands in bmad-output/; a fix story is drafted on handoff.
Auditable agent team: a BMAD fully automated R&D pipeline orchestrator. Orchestrates 9 specialized Agents (PM, Architect, UI Designer, Tech Lead, Scrum Master, Frontend, Backend, QA, DevOps) from requirements to deployment, with event sourcing + non-bypassable gates + deterministic evals at every step — verifiable that tests really run and gates really pass. Supports single-stage slice commands (/boss:plan /review /qa /ship) and a pure Markdown fallback without the CLI. Triggers (full pipeline): 'boss mode', '/boss', '全自动开发', '从需求到部署', '帮我做一个', 'build this', 'ship it', '全流程', '自动化开发', '一键开发', 'start a project', 'new feature' Triggers (single-stage slices, most commonly used on existing projects): - /boss:review — '帮我评审', '代码评审', '技术评审', '看看这段代码有什么问题', '风险评估', 'review this', 'review my PR', 'code review' - /boss:qa — '跑一下测试', '补测试', '检查质量', '测试覆盖够不够', '门禁过了吗', 'run the tests', 'check quality', 'add tests' - /boss:plan — '先出个方案', '写个 PRD', '设计一下架构', 'write a spec', 'plan this feature' - /boss:ship — '可以发布了吗', '部署检查', 'ready to ship', 'deploy checks' Does NOT trigger: - Single-file changes or simple bug fixes (just edit directly) - Pure code reading or explanation (use the read tool) - Repeated invocation while a pipeline is already running - Extremely small tasks (estimated <30 minutes of manual work, no PRD/architecture/gate records needed) Output: complete project code + PRD/architecture/UI/test/deployment documents, written to the .boss/<feature>/ directory
Use when building interactive terminal dashboards or full-screen terminal applications; when implementing keyboard navigation, live data updates, or multi-panel layouts; when TUI is flickering, slow, or unresponsive; when handling terminal resize events
Use the free built-in .NET SDK analyzers and analysis levels with gradual Roslyn warning promotion. USE FOR: the repo wants first-party .NET analyzers; CI should fail on analyzer warnings; the team needs AnalysisLevel or AnalysisMode guidance. DO NOT USE FOR: third-party analyzer selection by itself; formatting-only work. INVOKES: inspect the repository context, edit targeted files, and run relevant build, test, lint, or validation commands when changes are made.
Plan, run, and monitor Fireworks training from a coding agent. Covers managed SFT, DPO, ORPO, and RFT plus Training API serverless and dedicated workflows, complete cost and parameter confirmation, active monitoring, checkpoints, deployment, resume, and teardown. Use for explicit train, fine-tune, deploy, or resume requests. Use research when method, data, or evaluation is unclear; use debug for a stuck or failed run.
Convex backend guidance for queries, mutations, actions, schema, validators, auth, components, agents, search, testing, production config, and client/server integration. Use whenever interacting with a Convex backend, including Convex functions, schema changes, HTTP endpoints, scheduling, storage, AI features, or framework integration.
Main entry point for docs. Triggers: writing/updating developer guides, user guides, API references; does not include docs-neat final cleanup. Do not use for new feature implementation (cs-feat), bug fixing / error diagnosis (cs-issue), breaking down large requirements (cs-epic), domain modeling, or ADR decision records (cs-domain).
Install a complete design-token theme into any project in one pass -- pulls a style from the Open Agent Stack design-styles collection (aurora-mesh, cirrus, liquid-glass, mono-brutalist, neo-terminal, sand-terra, tidal) or a local tokens.json, wires it into Tailwind/CSS, and repaints existing components to match.
Use for every user-visible change in a GNOME app, new or existing, whether or not the request uses design vocabulary. Trigger on the repo rather than the wording — if it depends on gtk4/libadwaita (or is a GNOME-targeted Qt/PySide6 app), anything a user can see or interact with goes through this skill, including requests phrased purely as symptoms ("can I right-click to delete", "the window opens too small", "reordering doesn't work"). Covers choosing widgets, icons, dialogs, menus, sidebars, navigation, and empty states; interaction bugs in context menus, drag-and-drop reordering, window sizing, layout, duplicate affordances, and controls missing or inert at launch; and HIG review. Complements developing-gtk-apps, which owns architecture, threading, language syntax, and build/test plumbing.
Figma mastery and design-to-code: Auto Layout, component architecture, variable modes, Dev Mode handoff, Figma MCP server integration, Code Connect, and token pipelines. Use when working in Figma, converting designs to code, or setting up a design-to-code flywheel.
Provides comprehensive guidance for Flask framework including routing, templates, forms, database integration, extensions, and deployment. Use when the user asks about Flask, needs to create Flask applications, implement web routes, or build Python web applications.
Publish any Freshworks Platform 3.0 custom app via MCP tools: fdk validate/pack, app-upload, and submit/update through openai-server. Pre-publish: confirm Developer JWT matches manifest product modules (Freshdesk support_* vs Freshservice service_*; multiproduct sequential). At publish time, ask new vs existing listing; for new listings, prompt for supportEmail before create_app_upload_url (required for submit_custom_app). list_custom_apps for updates so the developer selects appId, then MCP handover (submit_custom_app or add_app_version with uploadId). Use when the user wants to push an app to the Marketplace (test), check publish status, or list existing apps. Pair with fw-app-dev for manifest or module fixes. Works with Cursor, Claude Code, and any MCP-compliant client.
Add, audit, or fix GitHub Copilot repository custom instructions — `.github/copilot-instructions.md` (repo-wide) and `.github/instructions/*.instructions.md` (path-specific) — so Copilot Chat, Copilot code review, and the Copilot coding agent all pick up project-wide guidance. Also covers `AGENTS.md`, `CLAUDE.md`, and `GEMINI.md` as alternate agent instruction files. Use when the user says "add Copilot instructions", "set up copilot-instructions.md", "audit our Copilot config", "make Copilot follow our conventions", "configure Copilot code review", "add path-specific Copilot rules", "applyTo glob", "excludeAgent", "Copilot is ignoring our guidance", or invokes /github-copilot-repo-instructions. Do NOT use for personal (per-user) Copilot instructions; those are out of scope for this skill. For AGENTS.md authoring guidance beyond file layout, see the agents-md skill.
Go engineering standards for this repo: package layout, error handling, logging, testing, concurrency, and design patterns. TRIGGER — load BEFORE writing or editing Go, not after: whenever any .go, go.mod, go.sum, or go.work file is touched; Go or Golang is named; or the working directory contains go.mod and the task is implementation, refactoring, review, or test writing. Plain feature requests in a Go repo (“add an endpoint”, “extract a helper”, “write a test for this”) count; the user will not say “idiomatic”. Load it even when another skill (tdd, simplify) is active. SKIP only when no Go file is involved and Go is not named.
Fast-track an urgent fix through a streamlined pipeline. Skips Product/Feature Council, applies the fix, runs a focused review, and creates a PR with optional Deployment Council. Use for production bugs, security patches, or critical regressions that cannot wait for the full planning pipeline.
Emulated Linear GraphQL API for local development and testing. Use when the user needs to test Linear integrations locally, emulate Linear issues, comments, teams, workflow states, OAuth apps, webhooks, agent sessions, or work with the Linear API without hitting the real Linear service. Triggers include "Linear API", "emulate Linear", "mock Linear", "test Linear OAuth", "Linear webhook", "Linear agent", "local Linear", or any task requiring a local Linear API.
Route MaaFramework factual questions to version-pinned primary sources—field semantics and defaults, parameter version support, API signatures, JSON schemas, protocol behavior, and version changes. Use for any MaaFW fact-check ("what does this field mean / what is the default / which version introduced it") or whenever another MaaFW skill says to "route MaaLLMWiki." Answers must cite pinned sources; never answer from memory. Use for any MaaFramework factual question—field semantics and defaults, version support, API signatures, schemas, protocol behavior and version changes—and whenever another MaaFW skill says to route through MaaLLMWiki. Answers must cite the version-pinned upstream source, never model memory.
Answer questions about Module Federation (MF) — configuration, runtime API, build plugins (Webpack/Rspack/Rsbuild/Vite), framework integration (React/Vue/Next.js/Modern.js/Angular), shared dependencies, exposes, remotes, debugging, troubleshooting, and micro-frontend architecture. Use this skill when the user asks anything about module federation, remote modules, shared deps, mf-manifest, federation runtime, or micro-frontends with MF.
nuqs reference (type-safe URL search params state management, React / Next.js). useQueryState, useQueryStates, parsers such as parseAsString / parseAsInteger / parseAsArrayOf, NuqsAdapter (Next.js / React Router / TanStack Router / Remix), createSerializer, createStandardSchemaV1, inferParserType, shallow routing, history control, SSR, URL length limits, and integration with Zod / Effect / Sury
Design, change, or review an HTTP/JSON API surface — endpoints, request/response shapes, authentication and authorization, pagination, idempotency, rate limits, versioning, and deprecations. Use when adding or modifying an HTTP endpoint, reviewing an OpenAPI spec or HTTP route diff, or deciding whether an HTTP API change breaks consumers. Do not use as a protocol-compatibility checklist for gRPC/protobuf, GraphQL, WebSockets, or other non-HTTP/JSON interfaces.
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.