MLOps
模型部署、评估与运维
verl-rl-training
Orchestra-Research
Provides guidance for training LLMs with reinforcement learning using verl (Volcano Engine RL). Use when implementing RLHF, GRPO, PPO, or other RL algorithms for LLM post-training at scale with flexible infrastructure backends.
llama-factory
Orchestra-Research
Expert guidance for fine-tuning LLMs with LLaMA-Factory - WebUI no-code, 100+ models, 2/3/4/5/6/8-bit QLoRA, multimodal support
deepspeed
Orchestra-Research
Expert guidance for distributed training with DeepSpeed - ZeRO optimization stages, pipeline parallelism, FP16/BF16/FP8, 1-bit Adam, sparse attention
axolotl
Orchestra-Research
Expert guidance for fine-tuning LLMs with Axolotl - YAML configs, 100+ models, LoRA/QLoRA, DPO/KTO/ORPO/GRPO, multimodal support
quantizing-models-bitsandbytes
Orchestra-Research
Quantizes LLMs to 8-bit or 4-bit for 50-75% memory reduction with minimal accuracy loss. Use when GPU memory is limited, need to fit larger models, or want faster inference. Supports INT8, NF4, FP4 formats, QLoRA training, and 8-bit optimizers. Works with HuggingFace Transformers.
mamba-architecture
Orchestra-Research
State-space model with O(n) complexity vs Transformers' O(n²). 5× faster inference, million-token sequences, no KV cache. Selective SSM with hardware-aware design. Mamba-1 (d_state=16) and Mamba-2 (d_state=128, multi-head). Models 130M-2.8B on HuggingFace.
causal-inference
brycewang-stanford
This skill covers causal inference methods in observational and quasi-experimental settings. Use when the user is implementing, choosing between, or debugging causal identification strategies — including instrumental variables, difference-in-differences, regression discontinuity, synthetic control, or matching estimators. Triggers on "causal effect", "identification strategy", "instrumental variable", "2SLS", "GMM", "difference-in-differences", "DiD", "staggered treatment", "regression discontinuity", "RDD", "synthetic control", "matching", "propensity score", "IPW", "AIPW", "doubly robust", "LATE", "ATT", "ATE", "parallel trends", "exclusion restriction", "first stage", "weak instruments", or "endogeneity".
pytorch-fsdp2
Orchestra-Research
Adds PyTorch FSDP2 (fully_shard) to training scripts with correct init, sharding, mixed precision/offload config, and distributed checkpointing. Use when models exceed single-GPU memory or when you need DTensor-based sharding with DeviceMesh.
huggingface-accelerate
NousResearch
Simplest distributed training API. 4 lines to add distributed support to any PyTorch script. Unified API for DeepSpeed/FSDP/Megatron/DDP. Automatic device placement, mixed precision (FP16/BF16/FP8). Interactive config, single launch command. HuggingFace ecosystem standard.
code-review-context
openai
Model visible context
cuml-machine-learning
langchain-ai
Use for GPU-accelerated machine learning on tabular data using NVIDIA cuML. Triggers when tasks involve classification, regression, clustering, dimensionality reduction, or model training on datasets.
game-hacking-techniques
gmh5225
Guide for game-hacking technique taxonomy and threat modeling relevant to game security. Use this skill when researching memory access, code injection, overlays, input simulation, engine-specific attack surfaces, or how modern anti-cheat systems constrain user-mode, kernel-mode, hypervisor, and DMA-based cheat implementations.
azure-ai-ml-py
microsoft
Azure Machine Learning SDK v2 for Python. Use for ML workspaces, jobs, models, datasets, compute, and pipelines. Triggers: "azure-ai-ml", "MLClient", "workspace", "model registry", "training jobs", "datasets".
azure-ai-projects-py
microsoft
Build AI applications using the Azure AI Projects Python SDK (azure-ai-projects). Use when working with Foundry project clients, creating versioned agents with PromptAgentDefinition, running evaluations, managing connections/deployments/datasets/indexes, or using OpenAI-compatible clients. This is the high-level Foundry SDK - for low-level agent operations, use azure-ai-agents-python skill.
azure-ai-contentunderstanding-py
microsoft
Azure AI Content Understanding SDK for Python. Use for multimodal content extraction from documents, images, audio, and video. Triggers: "azure-ai-contentunderstanding", "ContentUnderstandingClient", "multimodal analysis", "document extraction", "video analysis", "audio transcription".
nemoclaw-user-guide
NVIDIA
"Find official NemoClaw documentation for installation, configuration, operation, or troubleshooting. Also handles requested docs MCP setup."
jupyter-notebook
openai
"Use when the user asks to create, scaffold, or edit Jupyter notebooks (.ipynb) for experiments, explorations, or tutorials; prefer the bundled templates and run the helper script new_notebook.py to generate a clean starting notebook."
elizaos
elizaOS
"Use when the task involves elizaOS core runtime concepts, plugins, actions, providers, evaluators, services, memories, state composition, or upstream elizaOS development. Covers the main abstractions and the TypeScript runtime mental model."
effective-dbt-sql
lightdash
Use when writing or modifying dbt model SQL — deciding whether to add a subquery or reuse an existing dimension/metric, structuring a query, joining models, or refactoring a metric's SQL. Encodes SQL semantic-correctness rules: reuse existing fields, prefer CTEs over correlated subqueries, and make joins and column lists explicit.
python-development
apache
Guides Python SDK development in Apache Beam, including environment setup, testing, building, and running pipelines. Use when working with Python code in sdks/python/.
atlas-factor-change-protocol
kaecer68
"Mandatory 8-step protocol for adding, removing, or renaming FactorType values in atlas-go. Ensures optimizer constants, weight engine maps, domain structs, pipeline score structs, and event adjustments stay synchronized."
ib-write-tools
wavespestcontrolfl
Use when adding or changing an Intelligence Bar tool that writes anything (create, update, send, schedule) — every write goes through the confirm trust boundary and must register in write-gates.js.
ml-failfast-validation
terrylica
POC validation patterns to catch issues before committing to long-running ML experiments. TRIGGERS - fail-fast, POC validation, preflight check
DDD Domain Mapping (from QCSD-AGENTIC-QE-MAPPING-FRAMEWORK.md)
proffesor-for-testing
Agentic QE Fleet is an open-source AI-powered quality engineering platform designed for use with Claude Code, featuring specialized agents and skills to support testing activities for a product at any stage of the SDLC. Free to use, fork, build, and contribute. Based on the Agentic QE Framework created by Dragan Spiridonov.