Skills
97,905 packaged capabilities and playbooks for agents. Install any of them with one command.
Development
Guides fine-tuning of large language models with LLaMA-Factory, covering the WebUI no-code interface, training across 100+ supported models, quantized QLoRA at 2/3/4/5/6/8-bit precision, and multimodal model support. Use when setting up or configuring a LLaMA-Factory fine-tuning run, choosing a quantization level for QLoRA training, using the WebUI instead of writing training code, fine-tuning a multimodal model, or debugging a LLaMA-Factory training error. Not for general-purpose inference serving or for other training frameworks; use the matching skill for those.
Comprehensive logging and observability patterns for production systems including structured logging, distributed tracing, metrics collection, log aggregation, and alerting. Triggers for this skill - log, logging, logs, trace, tracing, traces, metrics, observability, OpenTelemetry, OTEL, Jaeger, Zipkin, structured logging, log level, debug, info, warn, error, fatal, correlation ID, span, spans, ELK, Elasticsearch, Loki, Datadog, Prometheus, Grafana, distributed tracing, log aggregation, alerting, monitoring, JSON logs, telemetry.
Driving a code change or document update through reviewed research, plan, and design outputs → approval → execution → verification → independent review, including blocked verification, pending review/CI, and run recovery. Owns phase placement, approval revision, evidence, finding state, and recovery meaning—not generic deliverable-review method, check-failure diagnosis, host execution APIs, or delivery storage. Development practices own verification diagnosis; conduct practices own non-code deliverable review; project policy chooses gates and branches. Defer to a host project's own more-specific change loop where present. Not for standalone read-only questions, investigations, or reviews that change nothing.
CRITICAL: Use for Makepad 2.0 app structure and Rust integration. Triggers on: makepad app, makepad getting started, app_main!, App::run, MatchEvent, AppMain, handle_event, handle_actions, ScriptVm, from_script_mod, makepad boilerplate, makepad new project, makepad cargo, Cargo.toml setup, hot reload, --hot, live reload, wasm deploy, cargo makepad, media plugin, audio_output, audio_input, AudioBuffer, cx.audio, makepad audio, 音频, 应用结构, 入门, 新项目, 脚手架, 启动, 热重载, 部署
CRITICAL: Use for Makepad 2.0 Splash scripting language. Triggers on: splash language, makepad script, script_mod!, makepad scripting, splash 脚本, makepad 2.0 script, mod.state, on_render, script_eval, streaming evaluation, splash syntax, splash vm, let binding, splash functions, hot reload, live reload, ScriptModKey, script_mod_overrides, checkpoint, incremental parsing, canvas splash, POST splash, fn tick, on_audio, set_text, tab switching, 音乐播放器, token monitor, driver script, audio API, 热重载, 脚本引擎, 增量解析
CRITICAL: Use for Makepad 2.0 troubleshooting and common mistakes. Triggers on: makepad error, makepad bug, makepad problem, makepad issue, makepad not working, text invisible, widget not showing, click not working, height zero, makepad pitfall, makepad gotcha, makepad FAQ, makepad help, script_mod error, compile error, widget not found, render not updating, hot reload not working, wasm build error, port conflict, server lock, IME popup, selection handle, popup window crash, canvas splash, POST splash loop, 100% CPU, set_visible not working, on_render empty, event bridge unreliable, float time display, fn tick not called, on_audio not called, button click through, 常见错误, 问题排查, 故障排除, 不显示, 不工作, 看不见, 热重载, 编译错误
CRITICAL: Use for Makepad 2.0 widget catalog and usage. Triggers on: makepad widget, makepad View, makepad Button, makepad Label, makepad TextInput, makepad PortalList, makepad Dock, makepad Modal, makepad Image, makepad CheckBox, makepad Slider, makepad DropDown, widget catalog, widget reference, widget list, SolidView, RoundedView, ScrollYView, FoldHeader, Splitter, FileTree, 组件, 控件, 视图, 按钮, 标签, 输入框, 列表, 模态框
Model Context Protocol (MCP) server development and tool management. Languages: Python, TypeScript. Capabilities: build MCP servers, integrate external APIs, discover/execute MCP tools, manage multi-server configs, design agent-centric tools. Actions: create, build, integrate, discover, execute, configure MCP servers/tools. Keywords: MCP, Model Context Protocol, MCP server, MCP tool, stdio transport, SSE transport, tool discovery, resource provider, prompt template, external API integration, Gemini CLI MCP, Claude MCP, agent tools, tool execution, server config. Use when: building MCP servers, integrating external APIs as MCP tools, discovering available MCP tools, executing MCP capabilities, configuring multi-server setups, designing tools for AI agents.
Build the actual product — from landing page to deployed MVP with payment integration, QA testing, and post-deploy canary monitoring. Handles code generation, deployment, database setup, authentication, Stripe/payment integration, systematic QA protocol, and production health scoring. Use when the user needs to build something, deploy a product, set up payments, create a landing page, or says 'build this', 'deploy', 'create MVP', 'set up payments', or 'ship it'.
Write MoonBit bindings to JavaScript with `extern "js"`. Use for FFI declarations against browser/Node/Deno APIs or npm packages, wrapping JS objects in opaque types, bridging Promises (`async fn` / `Promise::wait()`), `moon.pkg` exports (esm/cjs/iife), and null/undefined at the JS boundary.
Provides comprehensive guidance for generating MyBatis-Plus code including Entity, Mapper, Service, ServiceImpl, Controller, DTO, VO, BO and other related objects from database tables. Use ONLY when the user explicitly mentions MyBatis-Plus, mybatis-plus-generator, or wants to generate code using MyBatis-Plus framework. This skill automatically generates standard CRUD methods and custom methods based on user requirements for MyBatis-Plus projects. Supports MVC and DDD architectures, Java and Kotlin languages. Do NOT trigger for generic code generation, JPA/Hibernate, or other ORM frameworks.
Olares models via olares-cli router — Router (the AI gateway) and the Model Console inside a locally installed model application. Configure cloud vendors and their models, manage local LLM / embedding / audio / OCR model applications, edit a local model's card, name models with aliases, groups and default categories, issue API keys and quotas, read usage and audit, and call a model: chat, embed, rerank, search, scrape, translate, images, video, music, 3D, transcribe, speak, diarize, OCR. Requires Olares 1.12.7+. Use for Router, llm-gateway, AI gateway, 模型, 模型网关, 本地模型, add an OpenAI/Anthropic/DeepSeek key, install a Qwen or Gemma model, sk- key, model quota, token spend, model alias, default-chat, engine_args, which model answers by default, why a model call fails.
Build money-movement and fintech products on Polygon's Open Money Stack (OMS) API: neobanks, custodial wallets, remittance, on-ramps and off-ramps, stablecoin payments, and cash-in/cash-out. Use this whenever a developer wants to build, design, scope, or plan a financial product on Polygon, mentions the OMS API, asks how to authenticate or onboard with Polygon payments, says "build me a neobank" (or a remittance app, payments app, wallet, on-ramp, or off-ramp), or needs an implementation plan that maps their use case to real Polygon endpoints. The skill walks the developer through authentication, asks a few discovery questions, and produces an endpoint-grounded implementation plan that clearly separates what is callable today from what is early access.
Diagnose and fix the OpenViking memory plugin on this machine: the plugin install (enablement, hooks, MCP server), the client configuration (ovcli.conf / ov.conf / OPENVIKING_* env) and the connection to the OpenViking server (reachability, 401/403, /mcp). Use whenever memory "isn't working": no <openviking-context> block, empty recall, captures not landing, missing or failing MCP memory tools, 401/403, an offline statusline, right after installing/updating the plugin or switching servers/keys, or when the user asks for the plugin's status. Triggers: "memory not working", "check OpenViking", "plugin status", "记忆没生效", "插件状态", "连不上 OpenViking", "recall 为空", "401".
Use when a Datex Studio package (ComponentModule) has changed and its consumers must be updated and republished up the dependency graph. Recursively finds every package that references the changed one (directly or transitively), re-pins each to the new published version, and republishes — bottom-up, interactively, stopping at applications (which are reported as stale, never republished). Triggers: "propagate this package change", "update everything that references package X", "cascade republish", "bump and republish all dependents", "roll out the new version of <package> to its consumers", "who needs republishing after I published <package>".
Fine-tunes LLMs with Hugging Face PEFT, using LoRA, QLoRA, IA3, AdaLoRA, prefix tuning, and prompt tuning so that under 1% of parameters are trained. Covers rank, alpha, and target module selection, loading and merging adapters, multi-adapter serving, and integration with TRL SFTTrainer, Axolotl, and vLLM. Use when fine-tuning 7B-70B models on limited GPU memory, when running QLoRA on a single 24GB GPU, when choosing LoRA rank and alpha, when merging or swapping adapters on one base model, or when debugging CUDA OOM or an adapter that does not apply. Not for full fine-tuning of small models under 1B parameters or cases needing all weights updated.
Image-to-code replication pipeline. When the user provides a screenshot or design reference, this skill runs a structured extraction across seven layers (grid, type, color, spacing, components, atmosphere, interaction), builds an Extraction Sheet before any code is written, implements with exact fidelity to the reference, and verifies through inline Quality Gates and a final Visual Diff. The reference image is the spec. The code is a translation, not an interpretation.
Platform-specific building systems for mobile, VR, and accessibility. Use when implementing touch controls for building games, VR spatial input, colorblind-friendly feedback, or cross-platform building mechanics. Covers input adaptation and inclusive design patterns.
Monetizes browser games with Play.fun by guiding game registration, rewards and leaderboard setup, SDK integration, and deployment. Use when the user says "monetize my game", "add Play.fun", "add token rewards", "add a leaderboard", "register my game", "deploy so players earn rewards", or "launch a Playcoin". Do NOT use for generic game design, unrelated Solana questions, or non-game web apps.
This skill should be used when handling issue intake rotation duties for the Positron repository. It provides workflows for reviewing and organizing new issues, responding to discussions, handling support tickets, and searching for related content. Use this skill when on intake rotation duty, when helping someone with intake tasks, or when learning the intake rotation process.
Product Owner (/po, alias: Max, /max) - Senior PO with 10+ years agile experience. Use when defining product vision, creating/prioritizing backlog, writing user stories with acceptance criteria, making scope decisions, validating features against business goals, or planning releases and sprints.
Reverse-engineer the current state of analytics tracking from a codebase. Scans for SDK calls, identity management, and instrumentation patterns to produce a factual inventory — not recommendations. Outputs .telemetry/current-state.yaml and a timestamped audit report. Use when the user wants to know what's currently tracked, audit existing analytics, capture tracking reality before designing a new plan, 'what analytics do we have,' 'what events are tracked,' 'scan for tracking,' or 'tracking inventory.'
Manages session state and context handoffs for multi-session projects using the Session Handoff Protocol. Creates and maintains SESSION.md to track phase progress, git checkpoints, and next actions across context clears. Integrates with project-planning skill to convert IMPLEMENTATION_PHASES.md into trackable session state. Use when starting new projects after planning, resuming work after context clear, or managing complex multi-phase implementations. Keywords: session management, SESSION.md, session handoff protocol, context handoff, multi-session projects, phase tracking, git checkpoints, session state tracking, resume work, context clear, phase progress tracking, implementation phases, verification stage, debugging stage, next action tracking, work continuity, session recovery, context management, phased implementation tracking
Track progress across work sessions using SESSION.md with git checkpoints and concrete next actions. Converts IMPLEMENTATION_PHASES.md into trackable session state with phase status, progress markers, and recovery points. Use when: starting projects after planning phase, resuming work after context clears, managing multi-phase implementations, or troubleshooting lost context, missing progress tracking, or unclear next steps.
Publish completed work on a typed branch and optionally open a GitHub PR. Uses an approved draft with concrete QA steps and a high-level Mermaid Change diagram (Before/After for rework). Any agent that opens a PR follows the same ship contract, not only this skill. Use after work is complete when the user wants to ship a branch or PR. Never use inside /task.
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.