프레임워크 내부 구조

nnsight: Transparent Access to Neural Network Internals

Provides guidance for interpreting and manipulating neural network internals using nnsight with optional NDIF remote execution. Use when needing to run interpretability experiments on massive models (70B+) without local GPU resources, or when working with any PyTorch architecture.

John-Wang-08090

Ngrx Patterns

NgRx best practices for actions, reducers, selectors, and effects. Auto-invoked when working with NgRx store files.

mayeedwin0

Refactoring Simplifying Method Calls

Improve method interfaces by renaming, separating queries from modifiers, and introducing parameter objects. Trigger: When cleaning up method signatures, implementing CQRS, or reducing parameter lists.

johnnystefan0

Agency Terminal Integration Specialist

Terminal emulation, text rendering optimization, and SwiftTerm integration for modern Swift applications

augustoheiss0

Memory Safety Patterns

Implement memory-safe programming with RAII, ownership, smart pointers, and resource management across Rust, C++, and C. Use when writing safe systems code, managing resources, or preventing memory bugs.

Cpp Pro

Write idiomatic C++ code with modern features, RAII, smart pointers, and STL algorithms. Handles templates, move semantics, and performance optimization. Use PROACTIVELY for C++ refactoring, memory safety, or complex C++ patterns.

C Pro

Write efficient C code with proper memory management, pointer arithmetic, and system calls. Handles embedded systems, kernel modules, and performance-critical code. Use PROACTIVELY for C optimization, memory issues, or system programming.

Implement Plan

Implement a feature from a mini-plan document, user story, or GitHub issue using TDD, enforcing security and .NET/MCP best practices. Pass the path to a mini-plan .md file, user story, or GitHub issue URL/file. Also use when the user says "implement issue

MCrank0

Context Optimization

Apply compaction, masking, and caching strategies

Arm Cortex Expert

Senior embedded software engineer specializing in firmware and driver development for ARM Cortex-M microcontrollers (Teensy, STM32, nRF52, SAMD). Decades of experience writing reliable, optimized, and maintainable embedded code with deep expertise in memory barriers, DMA/cache coherency, interrupt-driven I/O, and peripheral drivers.

Implementer Hub

Trigger — "implement across multiple files", "coordinate implementation", "split work across modules", "implement this plan", "orchestrate the changes", "multi-file implementation". Skip — if only a single file needs changes (use `implementer-node-agent`), if context is not yet loaded (use `implementer-context` first), or if the plan does not exist yet (use `planner`).

Inference Kernel Optimization

Optimizes LLM inference kernels using FlashAttention, KV cache paging, Triton/CUDA kernel development, operator fusion, quantized GEMM, speculative decoding, and TensorRT-LLM. Use when reducing latency, memory, or throughput bottlenecks in model serving.

Llama Cpp

Compile and run llama.cpp for local inference — GGUF quantization, context sizing, and GPU offloading. Use when building llama.cpp from source, converting models to GGUF, configuring n_gpu_layers for partial offload, or tuning context size and batch parameters. Do not use for vLLM/TGI serving (prefer inference-serving) or model selection (prefer model-selection).

Purpose

Optimize LLM inference for CPU-only environments — quantization, threading, and memory mapping. Use when running models without GPU, optimizing llama.cpp for CPU, choosing quantization for RAM-constrained systems, or deploying inference on commodity hardware. Do not use for GPU inference (prefer inference-serving) or model selection (prefer model-selection).

Training Infrastructure

Configure distributed LLM training infrastructure—DDP, FSDP, DeepSpeed ZeRO, multi-node orchestration, checkpointing, fault tolerance, and mixed precision. Use when setting up torchrun/accelerate/deepspeed jobs, writing SLURM scripts, tuning NCCL, or debugging GPU memory and communication bottlenecks.

Purpose

Write and optimize Cloudflare Workers — implement fetch handlers, Durable Objects for stateful logic, service bindings, KV/R2 data access, scheduled/cron triggers, and middleware patterns with wrangler dev/deploy. Use when writing Worker code, debugging Worker runtime behavior, or designing Durable Object state machines. Do not use for Cloudflare DNS/WAF/Pages configuration (prefer cloudflare skill) or non-Worker serverless platforms.

Ai Principles

Use when writing, reviewing, or modifying AI/ML code using frameworks like OpenAI, Anthropic, LangChain, PyTorch, TensorFlow

asciifylabs0

Inject Docs

Inject framework-specific best practices and documentation into the project's CLAUDE.md file. Supports Next.js (via Vercel's agents-md codemod) and Python/FastAPI (via zhanymkanov/fastapi-best-practices). This skill should be used when a user wants to add framework documentation for AI coding agents or improve AI agent performance on framework-specific projects.

mgiovani0

Inject Nextjs Docs

Run the Next.js agents-md codemod to inject compressed framework documentation into the current project's CLAUDE.md or AGENTS.md. This skill should be used when a user wants to add Next.js framework docs to their project for AI coding agents, run the Vercel agents-md codemod, or improve AI agent performance on Next.js projects.

mgiovani0

Solidworks Macro

Génère des macros VBA pour SolidWorks 2021 avec intégration PDM

Ascenseurs-Menetrey-SA0

Idd Recipe Compilation

Compilation rules between IDD decompositions and Amplifier recipe YAML. Template expressions, foreach, while/break_when, parallel execution, approval gates, bidirectional compilation. Use when compiling decompositions to recipes or understanding the mapping.

michaeljabbour0

Optimize Size

Optimize Extended Min code size and branch speed by maximizing valid local fast branches/jumps while preserving compile correctness in both default and USE_ACCELERATOR builds.

michaelkamprath0

Hunting For Process Injection Techniques

Detect process injection techniques (T1055) including CreateRemoteThread, process hollowing, and DLL injection via Sysmon Event IDs 8 and 10 and EDR process telemetry

micsapp0

Path Tactics

Use ComputationalPaths path tactics to automate common RwEq goals (path_simp/path_auto/path_normalize), and structure calc-based proofs cleanly.

Arthur742Ramos0

Lean Hit Development

Guides adding new computational-path constructions to the ComputationalPaths library. Use when defining new spaces, π₁ calculations, encode-decode proofs, or adding new topological constructions.

Arthur742Ramos0

Add Cuda Kernel

Step-by-step tutorial for adding new CUDA kernels to FlashInfer

ariusewy0

Implement Ir Codegen

Implementar ou estender IR lowering e geração de código C do harbour-rust com foco em C legível, helpers explícitos e caminho executável.

arcostasi0

Tamagui Best Practices

Provides Tamagui patterns for config v4, compiler optimization, styled context, and cross-platform styling. Must use when working with Tamagui projects (tamagui.config.ts, @tamagui imports).

apostolos-geyer0