Total 59,052 skills
Showing 12 of 59052 skills
Implement Kubernetes security contexts, Pod Security Standards, and network policies. Secure cluster components and workloads. Use when hardening Kubernetes deployments or meeting security compliance.
Use before writing or editing any Python code — scripts, functions, bug fixes, cron jobs, data-processing one-offs, or additions to an existing .py file — even if the request sounds quick, casual, or "just a simple script." Applies whether the user names a .py file, describes what the script should do, or asks you to add/fix a function in Python. Covers required standards: type hints, no global variables, logging instead of print, wrapping any I/O (HTTP, DB, subprocess, file) in try/except, commenting the why not the what, writing a real test, and where to place design docs. Does not apply to conceptual Python questions with no code to write (e.g. explaining language features, comparing list vs tuple, explaining what an existing traceback means when the user hasn't asked for a fix, package install help, or learning-resource recommendations), and does not apply to non-Python languages.
在 Remotion 项目中创建带动画的金融 K 线图（蜡烛图）。涵盖数据获取（Yahoo Finance）、 插件架构（config / timeline / Scene / plugin）、Canvas 逐帧绘制、深色交易终端风格、 以及 Remotion Studio 预览和渲染输出。触发词：K线图、蜡烛图、candlestick、OHLC、 股票动画、金融图表视频、Remotion 图表。
One-time bootstrap of a persistent memory + self-evolution system for an agent workspace. Use when setting up memory for a new agent or workspace, or configuring dream/janitor crons.
Use for any codebase exploration or understanding task — reads historical intent from Entire checkpoints instead of guessing. Orchestrates other Entire skills to give the agent provenance-backed answers about code.
Use when you have an existing component, flow, or interface and need an evidence-backed accessibility design review after basic checks pass. Best for WCAG 2.2 compliance, focus management, ARIA pattern quality, semantics, and state communication gaps automated tools miss.
This skill should be used when the user asks to "add a sprite", "create a player", "add text to game", "create particles", "add a tilemap", "create a group", "add a container", "create game objects", "draw shapes", "add UI elements", "create a camera", "add depth", or needs to add any visual, interactive, or logical element to a Phaser 4 scene.
Autonomously optimizes end-to-end LLM inference throughput on AMD Instinct GPUs and reports a validated gain, using the Hyperloom multi-agent optimizer. Given a model, framework, workload (TP/EP, concurrency, ISL/OSL, precision), an objective and a time budget, it explores per-workload which levers to pull (serving/config parameters and env, framework enablement and source patches, and hot GPU-kernel rewrites), benchmarks each candidate, and returns the optimization stack that produced the gain. Use when the user wants to make a model serve faster, raise tokens/sec or throughput, optimize or tune vLLM or SGLang on MI300X/MI325X/MI355X, run Hyperloom, run the kernel-agent, quantize-then-optimize with Quark, set up Hyperloom from scratch, or resume a Hyperloom session. Do not use to stand up a server for plain serving, diagnose a broken ROCm install, or run a one-off kernel/benchmark or trace analysis without the optimization loop.
Serves an LLM on a supported AMD EPYC server CPU using vLLM with zentorch, in Docker, Podman, or conda. Use for "vLLM on CPU", "zentorch serving", or an EPYC CPU endpoint, including on a host that also has AMD Instinct GPUs. Detects the EPYC generation, validates the runtime, checks model support and RAM fit, sizes threads/KV/NUMA, confirms the plan, launches, and verifies the endpoint. Runs one instance on one socket and its memory. Reports and stops on failure; does not retry or debug. Use serving-llms-on-instinct when the endpoint should run on a GPU. Excludes multi-node, EPYC 4000, and pre-Zen4 EPYC without AVX-512.
Personal source engine for saved X bookmarks: ask, draft, decide, and plan from the user's own saves.
Lê o MVP.md gerado pelo MVPFy e transforma o contexto do produto em uma base de marca rastreável para estratégia, voz e assets.
(NS) Distinctive production UI — layout, typography, motion, and component polish that avoids generic AI-slop aesthetics. Use whenever the user builds or refines pages, components, dashboards, forms, or design-brief work, or asks for better UI/UX — even if they do not say "design". Load docs/context/design-brief.md when present. Do NOT use for backend-only work, requirements writing, or full SDD orchestration (use ns-spec-driven).