ai

Use when answering questions about Claude Code CLI features, configuration, hooks, skills, MCP, permissions, settings, sub-agents, teams, plugins, keybindings, IDE integration, CI/CD setup, troubleshooting, or any Claude Code usage topic.

Use when working with Apache Doris: table design, data models (Duplicate/Unique/Aggregate), partitioning, bucketing, SQL syntax, data import (Stream Load, Broker Load, INSERT INTO), data export, lakehouse (Hive/Iceberg/Hudi/Paimon catalogs), materialized views, query acceleration, inverted index, compute-storage decoupled mode, administration, or Doris ecosystem tools.

justfile (just) documentation — a command runner with make-inspired syntax. Covers recipes (parameters, dependencies, shebang/script recipes), justfile language (settings, strings, functions, constants, attributes, conditionals, backticks), modules/imports, variables, shell configuration, and CLI options. USE THIS SKILL WHEN the user asks about just, justfile syntax, just recipes, just modules, just functions, or just settings.

USE THIS SKILL WHEN working with Mastra (TypeScript AI agent framework): building agents, workflows, RAG pipelines, memory, voice, evals, MCP tools, observability/tracing, deploying to Mastra Cloud/Cloudflare/Vercel, or using Mastra client SDK. Triggers on: mastra, Mastra agent, Mastra workflow, createAgent, createWorkflow, mastra.run, MastraClient, @mastra/core.

Use when working with the OpenRouter API: model routing, provider selection, model variants (free/nitro/thinking), tool calling, structured outputs, prompt caching, OAuth, API keys, rate limits, streaming, embeddings, Claude Code integration, or any OpenRouter-specific feature.

Use when working with Ultralytics YOLO, including YOLO26/YOLO11/YOLOv8 models, detect/segment/classify/pose/OBB tasks, train/val/predict/export/track/benchmark modes, dataset YAML formats, HUB/platform workflows, integrations, solutions, or ultralytics Python API behavior.

Use when working with vLLM inference engine: OpenAI-compatible serving, model deployment, quantization (AWQ, GPTQ, FP8, GGUF, INT4/INT8), speculative decoding, LoRA adapters, structured outputs, tool calling, multimodal inputs, distributed serving (tensor/pipeline/expert/context parallel), Docker/Kubernetes deployment, engine configuration, memory optimization, PagedAttention, offline inference, CLI usage, or troubleshooting vLLM issues.

Use when answering questions about OpenAI platform features, guides, and concepts: Responses API usage, agents, function calling, text generation, vision, audio, embeddings, fine-tuning, batch processing, moderation, assistants, ChatKit, or any OpenAI developer guide topic.

Use when working with oRPC (OpenAPI Remote Procedure Call): defining type-safe RPC procedures, routers, middleware, context, error handling, event iterators (streaming), file upload/download, contract-first design, OpenAPI spec generation, server adapters (Next.js, Hono, Express, Fastify, Nuxt, Astro, Elysia, H3, SvelteKit, Remix, TanStack Start, Solid Start, WebSocket, Electron, Web Workers), client-side usage (RPC link, OpenAPI link, dynamic link), plugins (CORS, batch requests, compression, retry, CSRF, rate limit), integrations (TanStack Query, AI SDK, Better Auth, Sentry, OpenTelemetry, Pinia Colada, SWR), or migrating from tRPC.

Use when working with Ant Design React (`antd`) official docs in Chinese: getting started, ConfigProvider, theme tokens, CSS-in-JS, SSR, compatibility, i18n, migration to v6, CLI/MCP/LLMs.txt for agents, integrations with Vite/Next/Umi/Farm/Rsbuild/Refine, design spec/patterns, resources, and Ant Design architecture/blog guidance.

Use when working with the Anthropic REST API: Messages API, Batches API, Token Counting, Models API, Files API, Skills API, Admin API, rate limits, errors, versioning, service tiers, or supported regions. Triggers on anthropic api, messages api, api key, rate limit, batch processing, admin api.

Use when learning about Claude models, capabilities, pricing, prompt engineering, extended thinking, effort levels, vision, PDF support, streaming, structured outputs, tool use, MCP, context windows, prompt caching, embeddings, testing, guardrails, data residency, or Anthropic platform features.

Use when working with the Bun JavaScript runtime, package manager, bundler, or test runner: HTTP server, file I/O, SQLite, S3, Redis, shell scripting, FFI, workers, WebSocket, bunfig, bun install, bun test, bun build, bundler plugins, loaders, executables, macros, or any Bun API and configuration.

Use when working with ClickHouse: MergeTree engine family, SQL syntax, data types, table functions, materialized views, dictionaries, data modeling (denormalization, sharding, replication), integrations (Kafka, S3, PostgreSQL, MySQL), ClickHouse Cloud, Kubernetes operator, data compression, query optimization, or troubleshooting ClickHouse issues.

Use when working with Cloudflare AI and media products: Workers AI, AI Gateway, Agents, Agent Memory, AI Search, AI Crawl Control, Realtime, Stream, Images, Videos, model inference, agent workflows, media delivery, or AI observability.

Use when working with Cloudflare networking products: DNS, SSL/TLS, cache, load balancing, Magic Transit, Spectrum, BYOIP, network flow, interconnect, registrar, Web3, email routing/service, health checks, waiting room, or traffic routing.

Use when working with Cloudflare observability and account administration: Analytics, Logs, Logpush, Log Explorer, Radar, notifications, billing, support, Terraform/Pulumi, fundamentals, learning paths, reference architectures, account resources, or admin docs.

Use when working with Cloudflare Workers platform: Workers runtime APIs, Wrangler deployment, Pages, Durable Objects, Workflows, Containers, Browser Rendering, Sandbox, Workers VPC, bindings, routing, compatibility dates, or edge application troubleshooting.

Use when working with Cloudflare Zero Trust: Cloudflare One, Access, Gateway, WARP, Tunnel, device posture, DEX, CASB, Magic WAN/Cloudflare WAN, Zero Trust networking, 1.1.1.1, privacy proxy, or secure access troubleshooting.

ComfyUI documentation — node-based Stable Diffusion GUI and workflow engine. Covers 67 built-in nodes (CheckpointLoader, KSampler, CLIPTextEncode, VAEDecode, ControlNet, LoRA, IPAdapter, etc), custom node development (Python API, data types, UI widgets, backend extension), workflow tutorials (text-to-image, image-to-image, inpainting, ControlNet, LoRA, SDXL, Flux, upscaling, video, audio, 3D), ComfyUI Manager, Registry (publishing/versioning nodes), Comfy CLI, API reference, cloud deployment, installation (Desktop/manual/Docker), interface guide, and troubleshooting. USE THIS SKILL WHEN the user asks about ComfyUI nodes, workflows, custom node development, ComfyUI API, or Stable Diffusion pipeline configuration.

Gitea documentation — self-hosted Git service (lightweight GitHub/GitLab alternative) written in Go. Covers installation (Docker, binary, source, Kubernetes, cloud providers), administration (config cheat sheet, environment variables, reverse proxy, HTTPS, email, logging, authentication, backup/restore, Git LFS, GPG signing, fail2ban), usage (Actions CI/CD, packages registry for npm/PyPI/Docker/Maven/Go/Helm/Cargo/etc, webhooks, pull requests, AGit flow, protected branches/tags, repository mirroring, code owners, markdown), and development (API, OAuth2 provider, migrations, contributing guidelines). USE THIS SKILL WHEN the user asks about Gitea setup, configuration, Actions, package registry, API, or administration.

Use when administering GitLab Self-Managed or Dedicated: installation, configuration, Admin area settings, users, auth, Geo, Gitaly, PostgreSQL, Redis, Sidekiq, object storage, backups, monitoring, logs, upgrades, instance limits, feature flags, security, and troubleshooting.

Use when working with GitLab API docs: REST API endpoints, GraphQL API, authentication, pagination, rate limits, groups, projects, issues, merge requests, pipelines, jobs, packages, releases, users, access tokens, webhooks, admin APIs, and OpenAPI schemas.

Use when working with GitLab CI/CD docs: .gitlab-ci.yml syntax, jobs, stages, rules, workflow, include, inputs, components, variables, secrets, runners, executors, pipelines, artifacts, cache, environments, deployments, Docker, services, test reports, and pipeline troubleshooting.

Use when working with GitLab user-facing product docs: projects, groups, repositories, issues, merge requests, wikis, snippets, packages, releases, environments, application security, compliance, GitLab Duo, imports, search, permissions, profile, SSH, storage, and end-user workflows.

Use when working with the Gemini REST API: generateContent, streamGenerateContent, embedContent, caching, file uploads, token counting, model tuning, batch mode, live/real-time API, or any Google AI API endpoint parameters and response shapes.

Use when working with Grafana: dashboards, panels, visualizations, alerting, data sources (Prometheus, Loki, Elasticsearch, InfluxDB, MySQL, PostgreSQL, CloudWatch, Azure Monitor), provisioning, as-code (Terraform, Ansible, Grafana Operator), RBAC, SSO/SAML/LDAP, plugins, API, or Grafana setup and administration.

Use when working with Grafana plugins and plugin-specific docs, including data source/app/panel plugins such as Infinity, CloudWatch, ClickHouse, Zabbix, GitHub, GitLab, Jira, ServiceNow, MongoDB, OpenSearch, Volkov Labs panels, plugin configuration, query editors, variables, annotations, alerting, provisioning, or troubleshooting.

Use when working with Hugging Face Hub: model/dataset/Space repositories, Hub APIs, tokens, gated models/datasets, organizations, collections, cards, inference providers, jobs, storage/Xet, enterprise features, billing, webhooks, or Hub repository troubleshooting.

K3s documentation — lightweight Kubernetes for IoT/edge. Covers installation (air-gap, private registry, HA), architecture (embedded etcd, external datastore), networking (Flannel, CoreDNS, Traefik), security (secrets encryption, CIS hardening), upgrades, CLI (k3s server/agent/etcd-snapshot/certificate/token), and Helm charts. USE THIS SKILL WHEN the user asks about K3s, k3s server, k3s agent, k3s install, k3s networking, k3s hardening, or edge Kubernetes.

Use when working with llama.cpp: building from source, installation, GGUF model loading, quantization, speculative decoding, function calling, multimodal models (LLaVA, Gemma3, MiniCPM), GPU backends (CUDA, Metal, Vulkan, SYCL, OpenCL, CANN), Android deployment, Docker setup, performance tuning, or adding new model architectures to llama.cpp.

Use when working with OpenCode (anomalyco/opencode): terminal AI coding agent, configuration, model providers (Claude, OpenAI, Google, local models), custom agents, custom tools, MCP servers, LSP integration, key bindings, themes, modes, plugins, permissions, TUI, SDK, or OpenCode development setup.

Use when working with PyTorch APIs, tensors, autograd, torch.nn, torch.optim, torch.compile, export, distributed/RPC, CUDA/MPS/XPU backends, profiling, quantization, ONNX export, C++/LibTorch, or PyTorch runtime/debugging behavior.

Use when building, configuring, upgrading, testing, documenting, or publishing Storybook component workbenches, including CSF stories, main/preview configuration, framework integrations, addons, Doc Blocks, MDX, interaction/accessibility/visual tests, or Storybook MCP.

Use when working with Cloudflare security products: WAF, DDoS protection, Bot Management, API Shield, Turnstile, Rules, Ruleset Engine, firewall rules, network firewall, client-side security, challenges, DMARC, or security troubleshooting.

glab CLI documentation — GitLab's official command-line tool. Covers merge requests, issues, CI/CD pipelines, repositories, runners, releases, clusters, stacks, variables, schedules, labels, milestones, auth, config, snippets, deploy keys, SSH/GPG keys, Duo AI, MCP, and OpenTofu. USE THIS SKILL WHEN the user asks about glab commands, glab CLI reference, glab mr, glab ci, glab issue, glab repo, glab runner, or GitLab CLI docs.

Use when working with Linear product workflows, issues, projects, cycles, initiatives, views, integrations, workspace administration, AI agents, MCP, GraphQL API, OAuth, webhooks, or the Linear TypeScript SDK

USE THIS SKILL WHEN working with LLaMA Factory: fine-tuning LLMs (SFT/DPO/PPO/KTO/ORPO), LoRA/QLoRA training, dataset preparation (alpaca/sharegpt format), DeepSpeed/FSDP distributed training, LLaMA Board Web UI, model deployment (vLLM/SGLang), GRPO reinforcement learning, multimodal training, or using llamafactory-cli. Triggers on: LlamaFactory, llamafactory-cli, LLaMA Factory, llama-factory, LLaMA Board.

OpenObserve documentation — open-source observability platform for logs, metrics, traces, and RUM. Covers architecture, deployment, ingestion (OTLP, Fluent Bit, Vector, syslog), SQL functions, alerts, dashboards, pipelines, RBAC, API reference, and 50+ integrations (AWS, GCP, K8s, databases). USE THIS SKILL WHEN the user asks about OpenObserve setup, configuration, querying, alerting, or integration with observability pipelines.

Vite documentation — next-generation frontend build tool. Covers configuration (shared, server, build, preview, SSR, dep optimization, worker options), features (HMR, TypeScript, CSS/PostCSS, static assets, JSON, glob import, WebAssembly, Web Workers), plugin API (hooks, HMR API, environment API), SSR, backend integration, static deployment, dependency pre-bundling, env variables and modes, CLI, build optimization, migration guide (v5→v6→v7→v8), and Environment API (frameworks, instances, plugins, runtimes). USE THIS SKILL WHEN the user asks about Vite config, plugins, HMR, SSR, build optimization, or migration.