Guides performance investigation, benchmarking, and optimization of the Grida Canvas web editor (TypeScript reducer, Immer, React hooks). Use when profiling reducer dispatch cost, diagnosing slow interactions (drag, resize, color change), writing or running editor benchmarks, instrumenting with PerfObserver, or optimizing the JS-side state management pipeline.

Author SVG figures for Grida docs — diff-able, version-controlled vector diagrams embedded in doc pages instead of screenshots. Provides reusable primitives (selection chrome, size badges, anchor pins, resize cursors, click ripples), color/typography tokens, a starter template, and finished examples to crib from. Canvas user docs (docs/editor/) are the first consumer; the kit is meant to generalize to any product's docs. Use when drawing diagrams that explain UI behaviour — gestures, alignment, before/after states — that a screenshot alone can't capture. Trigger phrases: "svg diagram", "draw a figure", "visual for docs", "explain this gesture visually", "before/after diagram".

How to launch distributed Megatron-LM training jobs on a SLURM cluster. Covers a minimal sbatch skeleton, environment-variable setup for torch.distributed.run, CUDA_DEVICE_MAX_CONNECTIONS rules across hardware and parallelism modes, container conventions, monitoring, and per-rank failure diagnosis.

Use this skill when migrating applications, examples, integrations, documentation, manifests, or repository code from NeMo Flow to NeMo Relay across Python, Rust, Node.js, Go, C FFI, CLI, configuration, and observability surfaces.

Add a new cuTile GPU kernel operator to TileGym. Covers dispatch registration in ops.py, cuTile backend implementation, __init__.py exports, test creation, and benchmark in tests/benchmark. Use when adding, creating, or implementing a new cuTile operator/kernel in TileGym, or when asking how to register a new cuTile op.

Techniques for reducing peak GPU memory in Megatron Bridge — expandable segments, PEFT + SP input re-gather, parallelism resizing, activation recompute, CPU offloading constraints, and common OOM fixes.

Validate and use selective and full activation recompute in Megatron Bridge to reduce GPU memory usage at the cost of extra compute. Use for activation memory OOMs or regressions involving recompute_granularity, recompute_num_layers, recompute_modules, recompute_method, selective recompute, full recompute, or activation checkpointing.

章节归档 - 每章写完后的持久化操作，包括摘要生成、记忆更新和状态报告。关键词：归档、摘要、持久化、章节完成。

Prepare this repository for a release. Use when asked to do pre-release checks, summarize changes since a previous tag, update WHATSNEW.md or changelog entries, update CONTRIBUTORS.md, audit command long help or docs for release-visible features, or verify release documentation consistency.

Interactive Product Owner skill for requirements gathering, analysis, and PRD generation. Triggers when users request product requirements, feature specification, PRD creation, or need help understanding and documenting project requirements. Uses quality scoring and iterative dialogue to ensure comprehensive requirements before generating professional PRD documents.

How to update the bundled SQLite3 amalgamation (sqlite3/sqlite3.c and sqlite3.h), the Meson wrap, README.md, and CHANGELOG.md. Use when upgrading SQLite, refreshing the vendored amalgamation, or bumping the sqlite3 wrap.

SQLiteCpp release procedure: bump the version across all files, finalize the CHANGELOG, open the Release PR, and tag. Use when cutting a new release, bumping the project version, or tagging.

Writes high-quality unit tests for Kotlin Android codebases to increase meaningful test coverage. Use this skill to add tests for untested edge cases, ViewModel states, exception handling blocks, or UseCases using MockK, runTest, and the Given-When-Then pattern.

Optimizes Gradle build scripts, compilation times, and Android Studio sync performance. Use this skill to migrate kapt to ksp, fix Gradle Configuration Cache violations, extract build logic into precompiled script plugins, or tune gradle.properties for better JVM and parallel execution.

Elevates the Jetpack Compose user interface through micro-UX improvements, animations, accessibility (a11y) fixes, and UI structural polish. Use this skill to extract hardcoded styles to AppTheme, add TalkBack semantics, implement AnimatedVisibility or micro-animations, extract stateless Composables, and add @Preview annotations.

构建、扩展、审计或迁移可复用的 UI 组件与设计系统。当交付核心是跨页面或跨消费者复用的 token、组件 API、交互与无障碍合同、设计—代码映射、组件文档、分发包、版本治理或采用迁移时使用；若核心是完成一个产品工作流、阅读内容或单项视觉资产，则改用相应领域 Skill。

构建、重构、诊断和验证以玩家能动性为核心的二维游戏与趣味体验。用于二维规则/动作/益智游戏、互动玩具、叙事探索、节奏体验、生成体验和本地多人；按子型选择专业运行时与内容工具，分别验证机制、手感、内容、音频、可访问性、平台和玩家价值，且不以工程通过冒充好玩、公平或复玩价值。

Frequency-domain analysis — FFT, power spectral density (Welch/periodogram), spectrograms, wavelet transforms, and coherence. Use for any signal with periodic, quasi-periodic, or transient frequency content in physics data.

Update dotnet/macios to a new Xcode beta and validate it end-to-end. Use this skill when a user asks to bump Xcode beta versions, update macios SDK/version constants, run xtro-sharpie sanitization, and run introspection tests for iOS/tvOS/macOS/Mac Catalyst.

Researches a topic by breaking it into subtopics, gathering factual information with reasoning, and producing a structured summary with key findings and open questions. Use when the user asks to research, investigate, look up, summarize a topic, or says 'what is known about...' or 'learn about...'