模板
项目与代码模板
agent-tool-builder
davila7
"Tools are how AI agents interact with the world. A well-designed tool is the difference between an agent that works and one that hallucinates, fails silently, or costs 10x more tokens than necessary. This skill covers tool design from schema to error handling. JSON Schema best practices, description writing that actually helps the LLM, validation, and the emerging MCP standard that's becoming the lingua franca for AI tools. Key insight: Tool descriptions are more important than tool implementa"
training-llms-megatron
davila7
Trains large language models (2B-462B parameters) using NVIDIA Megatron-Core with advanced parallelism strategies. Use when training models >1B parameters, need maximum GPU efficiency (47% MFU on H100), or require tensor/pipeline/sequence/context/expert parallelism. Production-ready framework used for Nemotron, LLaMA, DeepSeek.
moe-training
davila7
Train Mixture of Experts (MoE) models using DeepSpeed or HuggingFace. Use when training large-scale models with limited compute (5× cost reduction vs dense models), implementing sparse architectures like Mixtral 8x7B or DeepSeek-V3, or scaling model capacity without proportional compute increase. Covers MoE architectures, routing mechanisms, load balancing, expert parallelism, and inference optimization.
nemo-curator
davila7
GPU-accelerated data curation for LLM training. Supports text/image/video/audio. Features fuzzy deduplication (16× faster), quality filtering (30+ heuristics), semantic deduplication, PII redaction, NSFW detection. Scales across GPUs with RAPIDS. Use for preparing high-quality training datasets, cleaning web data, or deduplicating large corpora.
dispatching-parallel-agents
davila7
Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies
llama-factory
davila7
Expert guidance for fine-tuning LLMs with LLaMA-Factory - WebUI no-code, 100+ models, 2/3/4/5/6/8-bit QLoRA, multimodal support
tensorrt-llm
davila7
Optimizes LLM inference with NVIDIA TensorRT for maximum throughput and lowest latency. Use for production deployment on NVIDIA GPUs (A100/H100), when you need 10-100x faster inference than PyTorch, or for serving models with quantization (FP8/INT4), in-flight batching, and multi-GPU scaling.
gepetto
davila7
Creates detailed, sectionized implementation plans through research, stakeholder interviews, and multi-LLM review. Use when planning features that need thorough pre-implementation analysis.
evaluating-llms-harness
davila7
Evaluates LLMs across 60+ academic benchmarks (MMLU, HumanEval, GSM8K, TruthfulQA, HellaSwag). Use when benchmarking model quality, comparing models, reporting academic results, or tracking training progress. Industry standard used by EleutherAI, HuggingFace, and major labs. Supports HuggingFace, vLLM, APIs.
datadog-cli
davila7
Datadog CLI for searching logs, querying metrics, tracing requests, and managing dashboards. Use this when debugging production issues or working with Datadog observability.
agent-manager-skill
davila7
Manage multiple local CLI agents via tmux sessions (start/stop/monitor/assign) with cron-friendly scheduling.
docs-search
davila7
Search auto-generated codebase documentation for function signatures, API docs, class definitions, and code comments. Use when the user asks to "search docs", "find documentation", "look up a function", "check the API", or before implementing changes to verify correct signatures and patterns.
harness-update
Chachamaru127
"Harnessを安全にアップデート。バックアップ付きで安心安全。Use when user mentions '/harness-update', update harness, update version, or template updates. Do NOT load for: app updates, dependency updates, product version bumps."
User Stories
product-on-purpose
A curated collection of 24 best-practice, plug-and-play product management “agent skills” plus templates and workflow bundles for consistent, professional PM outputs.
Solution Brief
product-on-purpose
A curated collection of 24 best-practice, plug-and-play product management “agent skills” plus templates and workflow bundles for consistent, professional PM outputs.
Problem Statement
product-on-purpose
A curated collection of 24 best-practice, plug-and-play product management “agent skills” plus templates and workflow bundles for consistent, professional PM outputs.
Interview Synthesis
product-on-purpose
A curated collection of 24 best-practice, plug-and-play product management “agent skills” plus templates and workflow bundles for consistent, professional PM outputs.
Stakeholder Summary
product-on-purpose
A curated collection of 24 best-practice, plug-and-play product management “agent skills” plus templates and workflow bundles for consistent, professional PM outputs.
Lessons Log
product-on-purpose
A curated collection of 24 best-practice, plug-and-play product management “agent skills” plus templates and workflow bundles for consistent, professional PM outputs.
scaffold-dotnet-test-project
dotnet
Create and wire the first .NET test project. USE FOR: "solution has no tests", Tests.csproj/ProjectReference, solution registration, central packages, or tests missing from CI. DO NOT USE when a suitable project exists, for migration, or for MSTest API/attribute/MSTest.Sdk/parallelization advice without a request to create and wire files (writing-mstest-tests).
plan-ui-change
dotnet
Plan complex Blazor UI features by decomposing them into focused components. USE FOR: building a complex Blazor page with multiple sections, planning component decomposition, designing a multi-section dashboard or layout, breaking down a large UI feature into composable components, pages with sidebars and content panels, any page with 3+ distinct visual sections or multiple interacting sub-features, identifying parent-child relationships and data flow. DO NOT USE FOR: creating new Blazor projects or apps from scratch (use create-blazor-project), implementing a single individual component (use author-component), writing component code with parameters and EventCallback (use author-component), or simple single-component pages.
maui-data-binding
dotnet
Guidance for .NET MAUI XAML and C# data bindings — compiled bindings, INotifyPropertyChanged / ObservableObject, value converters, binding modes, multi-binding, relative bindings, fallbacks, and MVVM best practices. USE FOR: setting up compiled bindings with x:DataType, implementing INotifyPropertyChanged or CommunityToolkit ObservableObject, creating IValueConverter / IMultiValueConverter, choosing binding modes, configuring BindingContext, relative bindings, binding fallbacks, StringFormat, code-behind SetBinding with lambdas, and enforcing XC0022/XC0025 warnings. DO NOT USE FOR: CollectionView item templates and layouts (use maui-collectionview), Shell navigation data passing (use maui-shell-navigation), dependency injection (use maui-dependency-injection), or animations triggered by property changes (use .NET MAUI animation APIs).
maui-collectionview
dotnet
Guidance for implementing CollectionView in .NET MAUI apps — data display, layouts (list & grid), selection, grouping, scrolling, empty views, templates, incremental loading, swipe actions, and pull-to-refresh. USE FOR: "CollectionView", "list view", "grid layout", "data template", "item template", "grouping", "pull to refresh", "incremental loading", "swipe actions", "empty view", "selection mode", "scroll to item", displaying scrollable data, replacing ListView. DO NOT USE FOR: simple static layouts without scrollable data (use Grid or StackLayout), map pin lists (use Microsoft.Maui.Controls.Maps), table-based data entry forms, non-MAUI list controls, CarouselView or BindableLayout questions, platform-specific handler or renderer customization, diagnosing CollectionView bugs in the MAUI framework itself, or general MVVM/binding questions that merely happen to mention a list (use maui-data-binding).
template-authoring
dotnet
Guides creation and validation of custom dotnet new templates from existing projects. Generates a .template.config/template.json that preserves the source project's conventions. USE FOR: creating a reusable dotnet new template from an existing project, bootstrapping .template.config/template.json with correct identity, shortName, parameters, and post-actions, adding parameters or conditional content to a template you are authoring, validating the template.json you are authoring before publishing, packaging templates as NuGet packages for distribution. DO NOT USE FOR: validating an existing template.json as a standalone task (use template-validation), finding or using existing templates (use template-discovery and template-instantiation), MSBuild project file issues unrelated to template authoring, NuGet package publishing (only template packaging structure).