Hugging Face
AICommonly used with
Skills using Hugging Face (326)
transformers
This skill should be used when working with pre-trained transformer models for natural language processing, computer vision, audio, or multimodal tasks. Use for text generation, classification, question answering, translation, summarization, image classification, object detection, speech recognition, and fine-tuning models on custom datasets.
audiocraft-audio-generation
PyTorch library for audio generation including text-to-music (MusicGen) and text-to-sound (AudioGen). Use when you need to generate music from text descriptions, create sound effects, or perform melody-conditioned music generation.
blip-2-vision-language
Vision-language pre-training framework bridging frozen image encoders and LLMs. Use when you need image captioning, visual question answering, image-text retrieval, or multimodal chat with state-of-the-art zero-shot performance.
distributed-llm-pretraining-torchtitan
Provides PyTorch-native distributed LLM pretraining using torchtitan with 4D parallelism (FSDP2, TP, PP, CP). Use when pretraining Llama 3.1, DeepSeek V3, or custom models at scale from 8 to 512+ GPUs with Float8, torch.compile, and distributed checkpointing.
evaluating-code-models
Evaluates code generation models across HumanEval, MBPP, MultiPL-E, and 15+ benchmarks with pass@k metrics. Use when benchmarking code models, comparing coding abilities, testing multi-language support, or measuring code generation quality. Industry standard from BigCode Project used by HuggingFace leaderboards.
evaluating-llms-harness
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.
fine-tuning-with-trl
Fine-tune LLMs using reinforcement learning with TRL - SFT for instruction tuning, DPO for preference alignment, PPO/GRPO for reward optimization, and reward model training. Use when need RLHF, align model with preferences, or train from human feedback. Works with HuggingFace Transformers.
gguf-quantization
GGUF format and llama.cpp quantization for efficient CPU/GPU inference. Use when deploying models on consumer hardware, Apple Silicon, or when needing flexible quantization from 2-8 bit without GPU requirements.
gptq
Post-training 4-bit quantization for LLMs with minimal accuracy loss. Use for deploying large models (70B, 405B) on consumer GPUs, when you need 4× memory reduction with <2% perplexity degradation, or for faster inference (3-4× speedup) vs FP16. Integrates with transformers and PEFT for QLoRA fine-tuning.
guidance
Control LLM output with regex and grammars, guarantee valid JSON/XML/code generation, enforce structured formats, and build multi-step workflows with Guidance - Microsoft Research's constrained generation framework
hqq-quantization
Half-Quadratic Quantization for LLMs without calibration data. Use when quantizing models to 4/3/2-bit precision without needing calibration datasets, for fast quantization workflows, or when deploying with vLLM or HuggingFace Transformers.
huggingface-accelerate
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.
huggingface-tokenizers
Fast tokenizers optimized for research and production. Rust-based implementation tokenizes 1GB in <20 seconds. Supports BPE, WordPiece, and Unigram algorithms. Train custom vocabularies, track alignments, handle padding/truncation. Integrates seamlessly with transformers. Use when you need high-performance tokenization or custom tokenizer training.
llama-cpp
Runs LLM inference on CPU, Apple Silicon, and consumer GPUs without NVIDIA hardware. Use for edge deployment, M1/M2/M3 Macs, AMD/Intel GPUs, or when CUDA is unavailable. Supports GGUF quantization (1.5-8 bit) for reduced memory and 4-10× speedup vs PyTorch on CPU.
llamaguard
Meta's 7-8B specialized moderation model for LLM input/output filtering. 6 safety categories - violence/hate, sexual content, weapons, substances, self-harm, criminal planning. 94-95% accuracy. Deploy with vLLM, HuggingFace, Sagemaker. Integrates with NeMo Guardrails.
long-context
Extend context windows of transformer models using RoPE, YaRN, ALiBi, and position interpolation techniques. Use when processing long documents (32k-128k+ tokens), extending pre-trained models beyond original context limits, or implementing efficient positional encodings. Covers rotary embeddings, attention biases, interpolation methods, and extrapolation strategies for LLMs.
mamba-architecture
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.
ml-engineer
Build production ML systems with PyTorch 2.x, TensorFlow, and modern ML frameworks. Implements model serving, feature engineering, A/B testing, and monitoring.
model-merging
Merge multiple fine-tuned models using mergekit to combine capabilities without retraining. Use when creating specialized models by blending domain-specific expertise (math + coding + chat), improving performance beyond single models, or experimenting rapidly with model variants. Covers SLERP, TIES-Merging, DARE, Task Arithmetic, linear merging, and production deployment strategies.
moe-training
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.
outlines
Guarantee valid JSON/XML/code structure during generation, use Pydantic models for type-safe outputs, support local models (Transformers, vLLM), and maximize inference speed with Outlines - dottxt.ai's structured generation library
peft-fine-tuning
Parameter-efficient fine-tuning for LLMs using LoRA, QLoRA, and 25+ methods. Use when fine-tuning large models (7B-70B) with limited GPU memory, when you need to train <1% of parameters with minimal accuracy loss, or for multi-adapter serving. HuggingFace's official library integrated with transformers ecosystem.
pyvene-interventions
Provides guidance for performing causal interventions on PyTorch models using pyvene's declarative intervention framework. Use when conducting causal tracing, activation patching, interchange intervention training, or testing causal hypotheses about model behavior.
quantizing-models-bitsandbytes
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.
segment-anything-model
Foundation model for image segmentation with zero-shot transfer. Use when you need to segment any object in images using points, boxes, or masks as prompts, or automatically generate all object masks in an image.
sentencepiece
Language-independent tokenizer treating text as raw Unicode. Supports BPE and Unigram algorithms. Fast (50k sentences/sec), lightweight (6MB memory), deterministic vocabulary. Used by T5, ALBERT, XLNet, mBART. Train on raw text without pre-tokenization. Use when you need multilingual support, CJK languages, or reproducible tokenization.
simpo-training
Simple Preference Optimization for LLM alignment. Reference-free alternative to DPO with better performance (+6.4 points on AlpacaEval 2.0). No reference model needed, more efficient than DPO. Use for preference alignment when want simpler, faster training than DPO/PPO.
stable-diffusion-image-generation
State-of-the-art text-to-image generation with Stable Diffusion models via HuggingFace Diffusers. Use when generating images from text prompts, performing image-to-image translation, inpainting, or building custom diffusion pipelines.
verl-rl-training
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.
coreweave-core-workflow-a
Deploy KServe InferenceService on CoreWeave with autoscaling and GPU scheduling. Use when serving ML models with KServe, configuring scale-to-zero, or deploying production inference endpoints on CoreWeave. Trigger with phrases like "coreweave inference service", "coreweave kserve", "coreweave model serving", "deploy model on coreweave".
coreweave-data-handling
Handle training data and model artifacts on CoreWeave persistent storage. Use when managing large datasets, configuring storage classes, or implementing data pipelines for GPU workloads. Trigger with phrases like "coreweave data", "coreweave storage", "coreweave pvc", "coreweave dataset management".
coreweave-security-basics
Secure CoreWeave deployments with RBAC, network policies, and secrets management. Use when hardening GPU workloads, managing model access, or configuring namespace isolation. Trigger with phrases like "coreweave security", "coreweave rbac", "secure coreweave", "coreweave secrets".
cortex-recon
ML reconnaissance — inventory all models, pipelines, data sources, and monitoring. Use when asked "what ML do we have", "model inventory", or "ML assessment".
exa-core-workflow-b
Execute Exa findSimilar, getContents, answer, and streaming answer workflows. Use when finding pages similar to a URL, retrieving content for known URLs, or getting AI-generated answers with citations. Trigger with phrases like "exa find similar", "exa get contents", "exa answer", "exa similarity search", "findSimilarAndContents".
hugging-face-jobs
Run workloads on Hugging Face Jobs with managed CPUs, GPUs, TPUs, secrets, and Hub persistence.
hugging-face-model-trainer
Train or fine-tune TRL language models on Hugging Face Jobs, including SFT, DPO, GRPO, and GGUF export.
ml-engineer
Build production ML systems with PyTorch 2.x, TensorFlow, and modern ML frameworks. Implements model serving, feature engineering, A/B testing, and monitoring.
transformers-js
Run Hugging Face models in JavaScript or TypeScript with Transformers.js in Node.js or the browser.
evaluating-llms-harness
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.
huggingface-accelerate
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.
ml-training-recipes
Battle-tested PyTorch training recipes for all domains — LLMs, vision, diffusion, medical imaging, protein/drug discovery, spatial omics, genomics. Covers training loops, optimizer selection (AdamW, Muon), LR scheduling, mixed precision, debugging, and systematic experimentation. Use when training or fine-tuning neural networks, debugging loss spikes or OOM, choosing architectures, or optimizing GPU throughput.
peft-fine-tuning
Parameter-efficient fine-tuning for LLMs using LoRA, QLoRA, and 25+ methods. Use when fine-tuning large models (7B-70B) with limited GPU memory, when you need to train <1% of parameters with minimal accuracy loss, or for multi-adapter serving. HuggingFace's official library integrated with transformers ecosystem.
ray-data
Scalable data processing for ML workloads. Streaming execution across CPU/GPU, supports Parquet/CSV/JSON/images. Integrates with Ray Train, PyTorch, TensorFlow. Scales from single machine to 100s of nodes. Use for batch inference, data preprocessing, multi-modal data loading, or distributed ETL pipelines.
evaluating-code-models
Evaluates code generation models across HumanEval, MBPP, MultiPL-E, and 15+ benchmarks with pass@k metrics. Use when benchmarking code models, comparing coding abilities, testing multi-language support, or measuring code generation quality. Industry standard from BigCode Project used by HuggingFace leaderboards.
evaluating-llms-harness
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.
coreweave-incident-runbook
Incident response runbook for CoreWeave GPU workload failures. Use when inference services are down, GPUs are unavailable, or responding to production incidents on CoreWeave. Trigger with phrases like "coreweave incident", "coreweave outage", "coreweave runbook", "coreweave service down".
detecting-ai-model-prompt-injection-attacks
Detects prompt injection attacks targeting LLM-based applications using a multi-layered defense combining regex pattern matching for known attack signatures, heuristic scoring for structural anomalies, and transformer-based classification with DeBERTa models. The detector analyzes user inputs before they reach the LLM, flagging direct injections (system prompt overrides, role-play escapes, instruction hijacking) and indirect injections (encoded payloads, multi-language obfuscation, delimiter-based escapes). Based on the OWASP LLM Top 10 (LLM01:2025 Prompt Injection) and Simon Willison's prompt injection taxonomy. Activates for requests involving prompt injection detection, LLM input sanitization, AI security scanning, or prompt attack classification.
esm
Use when working directly with the `esm` Python SDK, ESM3 or ESMC model IDs, Forge/Biohub inference clients, or ESMFold2 folding workflows.
hugging-science
Use when the user is doing AI/ML work in a scientific domain such as biology, chemistry, physics, astronomy, climate, genomics, materials, medicine, ecology, energy, engineering, math, drug discovery, protein design, weather modeling, theorem proving, single-cell, or PDE solving. Hugging Science is a curated catalog of scientific datasets, models, blog posts, and interactive Spaces. This skill helps discover and use resources via `datasets`, `transformers`, the HF Inference API, `gradio_client`, and methodology citations.
hugging-face-cli
Use the Hugging Face Hub CLI (`hf`) to download, upload, and manage models, datasets, and Spaces.
hugging-face-community-evals
Run local evaluations for Hugging Face Hub models with inspect-ai or lighteval.
hugging-face-dataset-viewer
Query Hugging Face datasets through the Dataset Viewer API for splits, rows, search, filters, and parquet links.
hugging-face-datasets
Create and manage datasets on Hugging Face Hub. Supports initializing repos, defining configs/system prompts, streaming row updates, and SQL-based dataset querying/transformation. Designed to work alongside HF MCP server for comprehensive dataset workflows.
hugging-face-evaluation
Add and manage evaluation results in Hugging Face model cards. Supports extracting eval tables from README content, importing scores from Artificial Analysis API, and running custom model evaluations with vLLM/lighteval. Works with the model-index metadata format.
hugging-face-paper-publisher
Publish and manage research papers on Hugging Face Hub. Supports creating paper pages, linking papers to models/datasets, claiming authorship, and generating professional markdown-based research articles.
hugging-face-papers
Read and analyze Hugging Face paper pages or arXiv papers with markdown and papers API metadata.
hugging-face-tool-builder
Your purpose is now is to create reusable command line scripts and utilities for using the Hugging Face API, allowing chaining, piping and intermediate processing where helpful. You can access the API directly, as well as use the hf command line tool.
hugging-face-trackio
Track ML experiments with Trackio using Python logging, alerts, and CLI metric retrieval.
hugging-face-vision-trainer
Train or fine-tune vision models on Hugging Face Jobs for detection, classification, and SAM or SAM2 segmentation.
data-formats
Working with diverse data formats: binary, text, structured, and custom
dataset-transformation
Generates a Jupyter notebook that transforms datasets between ML schemas for model training or evaluation. Use when the user says "transform", "convert", "reformat", "change the format", or when a dataset's schema needs to change to match the target format — always use this skill for format changes rather than writing inline transformation code. Supports OpenAI chat, SageMaker SFT/DPO/RLVR, HuggingFace preference, Bedrock Nova, VERL, and custom JSONL formats from local files or S3.
fine-tuning-expert
Use when fine-tuning LLMs, training custom models, or adapting foundation models for specific tasks. Invoke for configuring LoRA/QLoRA adapters, preparing JSONL training datasets, setting hyperparameters for fine-tuning runs, adapter training, transfer learning, finetuning with Hugging Face PEFT, OpenAI fine-tuning, instruction tuning, RLHF, DPO, or quantizing and deploying fine-tuned models. Trigger terms include: LoRA, QLoRA, PEFT, finetuning, fine-tuning, adapter tuning, LLM training, model training, custom model.
academic-plotting
Generates publication-quality figures for ML papers from research context. Given a paper section or description, extracts system components and relationships to generate architecture diagrams via Gemini. Given experiment results or data, auto-selects chart type and generates data-driven figures via matplotlib/seaborn. Use when creating any figure for a conference paper.
audiocraft-audio-generation
PyTorch library for audio generation including text-to-music (MusicGen) and text-to-sound (AudioGen). Use when you need to generate music from text descriptions, create sound effects, or perform melody-conditioned music generation.
autogpt-agents
Autonomous AI agent platform for building and deploying continuous agents. Use when creating visual workflow agents, deploying persistent autonomous agents, or building complex multi-step AI automation systems.
autoresearch
Orchestrates end-to-end autonomous AI research projects using a two-loop architecture. The inner loop runs rapid experiment iterations with clear optimization targets. The outer loop synthesizes results, identifies patterns, and steers research direction. Routes to domain-specific skills for execution, supports continuous agent operation via Claude Code /loop and OpenClaw heartbeat, and produces research presentations and papers. Use when starting a research project, running autonomous experiments, or managing a multi-hypothesis research effort.
awq-quantization
Activation-aware weight quantization for 4-bit LLM compression with 3x speedup and minimal accuracy loss. Use when deploying large models (7B-70B) on limited GPU memory, when you need faster inference than GPTQ with better accuracy preservation, or for instruction-tuned and multimodal models. MLSys 2024 Best Paper Award winner.
axolotl
Expert guidance for fine-tuning LLMs with Axolotl - YAML configs, 100+ models, LoRA/QLoRA, DPO/KTO/ORPO/GRPO, multimodal support
blip-2-vision-language
Vision-language pre-training framework bridging frozen image encoders and LLMs. Use when you need image captioning, visual question answering, image-text retrieval, or multimodal chat with state-of-the-art zero-shot performance.
brainstorming-research-ideas
Guides researchers through structured ideation frameworks to discover high-impact research directions. Use when exploring new problem spaces, pivoting between projects, or seeking novel angles on existing work.
chroma
Open-source embedding database for AI applications. Store embeddings and metadata, perform vector and full-text search, filter by metadata. Simple 4-function API. Scales from notebooks to production clusters. Use for semantic search, RAG applications, or document retrieval. Best for local development and open-source projects.
clip
OpenAI's model connecting vision and language. Enables zero-shot image classification, image-text matching, and cross-modal retrieval. Trained on 400M image-text pairs. Use for image search, content moderation, or vision-language tasks without fine-tuning. Best for general-purpose image understanding.
constitutional-ai
Anthropic's method for training harmless AI through self-improvement. Two-phase approach - supervised learning with self-critique/revision, then RLAIF (RL from AI Feedback). Use for safety alignment, reducing harmful outputs without human labels. Powers Claude's safety system.
creative-thinking-for-research
Applies cognitive science frameworks for creative thinking to CS and AI research ideation. Use when seeking genuinely novel research directions by leveraging combinatorial creativity, analogical reasoning, constraint manipulation, and other empirically grounded creative strategies.
crewai-multi-agent
Multi-agent orchestration framework for autonomous AI collaboration. Use when building teams of specialized agents working together on complex tasks, when you need role-based agent collaboration with memory, or for production workflows requiring sequential/hierarchical execution. Built without LangChain dependencies for lean, fast execution.
deepspeed
Expert guidance for distributed training with DeepSpeed - ZeRO optimization stages, pipeline parallelism, FP16/BF16/FP8, 1-bit Adam, sparse attention
distributed-llm-pretraining-torchtitan
Provides PyTorch-native distributed LLM pretraining using torchtitan with 4D parallelism (FSDP2, TP, PP, CP). Use when pretraining Llama 3.1, DeepSeek V3, or custom models at scale from 8 to 512+ GPUs with Float8, torch.compile, and distributed checkpointing.
dspy
Build complex AI systems with declarative programming, optimize prompts automatically, create modular RAG systems and agents with DSPy - Stanford NLP's framework for systematic LM programming
evaluating-code-models
Evaluates code generation models across HumanEval, MBPP, MultiPL-E, and 15+ benchmarks with pass@k metrics. Use when benchmarking code models, comparing coding abilities, testing multi-language support, or measuring code generation quality. Industry standard from BigCode Project used by HuggingFace leaderboards.
evaluating-cosmos-policy
Evaluates NVIDIA Cosmos Policy on LIBERO and RoboCasa simulation environments. Use when setting up cosmos-policy for robot manipulation evaluation, running headless GPU evaluations with EGL rendering, or profiling inference latency on cluster or local GPU machines.
evaluating-llms-harness
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.
experiment-tracking-swanlab
Provides guidance for experiment tracking with SwanLab. Use when you need open-source run tracking, local or self-hosted dashboards, and lightweight media logging for ML workflows.
faiss
Facebook's library for efficient similarity search and clustering of dense vectors. Supports billions of vectors, GPU acceleration, and various index types (Flat, IVF, HNSW). Use for fast k-NN search, large-scale vector retrieval, or when you need pure similarity search without metadata. Best for high-performance applications.
fine-tuning-openvla-oft
Fine-tunes and evaluates OpenVLA-OFT and OpenVLA-OFT+ policies for robot action generation with continuous action heads, LoRA adaptation, and FiLM conditioning on LIBERO simulation and ALOHA real-world setups. Use when reproducing OpenVLA-OFT paper results, training custom VLA action heads (L1 or diffusion), deploying server-client inference for ALOHA, or debugging normalization, LoRA merge, and cross-GPU issues.
fine-tuning-serving-openpi
Fine-tune and serve Physical Intelligence OpenPI models (pi0, pi0-fast, pi0.5) using JAX or PyTorch backends for robot policy inference across ALOHA, DROID, and LIBERO environments. Use when adapting pi0 models to custom datasets, converting JAX checkpoints to PyTorch, running policy inference servers, or debugging norm stats and GPU memory issues.
fine-tuning-with-trl
Fine-tune LLMs using reinforcement learning with TRL - SFT for instruction tuning, DPO for preference alignment, PPO/GRPO for reward optimization, and reward model training. Use when need RLHF, align model with preferences, or train from human feedback. Works with HuggingFace Transformers.
gguf-quantization
GGUF format and llama.cpp quantization for efficient CPU/GPU inference. Use when deploying models on consumer hardware, Apple Silicon, or when needing flexible quantization from 2-8 bit without GPU requirements.
gptq
Post-training 4-bit quantization for LLMs with minimal accuracy loss. Use for deploying large models (70B, 405B) on consumer GPUs, when you need 4× memory reduction with <2% perplexity degradation, or for faster inference (3-4× speedup) vs FP16. Integrates with transformers and PEFT for QLoRA fine-tuning.
grpo-rl-training
Expert guidance for GRPO/RL fine-tuning with TRL for reasoning and task-specific model training
guidance
Control LLM output with regex and grammars, guarantee valid JSON/XML/code generation, enforce structured formats, and build multi-step workflows with Guidance - Microsoft Research's constrained generation framework
hqq-quantization
Half-Quadratic Quantization for LLMs without calibration data. Use when quantizing models to 4/3/2-bit precision without needing calibration datasets, for fast quantization workflows, or when deploying with vLLM or HuggingFace Transformers.
huggingface-accelerate
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.
huggingface-tokenizers
Fast tokenizers optimized for research and production. Rust-based implementation tokenizes 1GB in <20 seconds. Supports BPE, WordPiece, and Unigram algorithms. Train custom vocabularies, track alignments, handle padding/truncation. Integrates seamlessly with transformers. Use when you need high-performance tokenization or custom tokenizer training.
implementing-llms-litgpt
Implements and trains LLMs using Lightning AI's LitGPT with 20+ pretrained architectures (Llama, Gemma, Phi, Qwen, Mistral). Use when need clean model implementations, educational understanding of architectures, or production fine-tuning with LoRA/QLoRA. Single-file implementations, no abstraction layers.
instructor
Extract structured data from LLM responses with Pydantic validation, retry failed extractions automatically, parse complex JSON with type safety, and stream partial results with Instructor - battle-tested structured output library
knowledge-distillation
Compress large language models using knowledge distillation from teacher to student models. Use when deploying smaller models with retained performance, transferring GPT-4 capabilities to open-source models, or reducing inference costs. Covers temperature scaling, soft targets, reverse KLD, logit distillation, and MiniLLM training strategies.
lambda-labs-gpu-cloud
Reserved and on-demand GPU cloud instances for ML training and inference. Use when you need dedicated GPU instances with simple SSH access, persistent filesystems, or high-performance multi-node clusters for large-scale training.
langchain
Framework for building LLM-powered applications with agents, chains, and RAG. Supports multiple providers (OpenAI, Anthropic, Google), 500+ integrations, ReAct agents, tool calling, memory management, and vector store retrieval. Use for building chatbots, question-answering systems, autonomous agents, or RAG applications. Best for rapid prototyping and production deployments.
langsmith-observability
LLM observability platform for tracing, evaluation, and monitoring. Use when debugging LLM applications, evaluating model outputs against datasets, monitoring production systems, or building systematic testing pipelines for AI applications.
llama-cpp
Runs LLM inference on CPU, Apple Silicon, and consumer GPUs without NVIDIA hardware. Use for edge deployment, M1/M2/M3 Macs, AMD/Intel GPUs, or when CUDA is unavailable. Supports GGUF quantization (1.5-8 bit) for reduced memory and 4-10× speedup vs PyTorch on CPU.
llama-factory
Expert guidance for fine-tuning LLMs with LLaMA-Factory - WebUI no-code, 100+ models, 2/3/4/5/6/8-bit QLoRA, multimodal support
llamaguard
Meta's 7-8B specialized moderation model for LLM input/output filtering. 6 safety categories - violence/hate, sexual content, weapons, substances, self-harm, criminal planning. 94-95% accuracy. Deploy with vLLM, HuggingFace, Sagemaker. Integrates with NeMo Guardrails.
llamaindex
Data framework for building LLM applications with RAG. Specializes in document ingestion (300+ connectors), indexing, and querying. Features vector indices, query engines, agents, and multi-modal support. Use for document Q&A, chatbots, knowledge retrieval, or building RAG pipelines. Best for data-centric LLM applications.
llava
Large Language and Vision Assistant. Enables visual instruction tuning and image-based conversations. Combines CLIP vision encoder with Vicuna/LLaMA language models. Supports multi-turn image chat, visual question answering, and instruction following. Use for vision-language chatbots or image understanding tasks. Best for conversational image analysis.
long-context
Extend context windows of transformer models using RoPE, YaRN, ALiBi, and position interpolation techniques. Use when processing long documents (32k-128k+ tokens), extending pre-trained models beyond original context limits, or implementing efficient positional encodings. Covers rotary embeddings, attention biases, interpolation methods, and extrapolation strategies for LLMs.
mamba-architecture
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.
miles-rl-training
Provides guidance for enterprise-grade RL training using miles, a production-ready fork of slime. Use when training large MoE models with FP8/INT4, needing train-inference alignment, or requiring speculative RL for maximum throughput.
ml-paper-writing
Write publication-ready ML/AI/Systems papers for NeurIPS, ICML, ICLR, ACL, AAAI, COLM, OSDI, NSDI, ASPLOS, SOSP. Use when drafting papers from research repos, structuring arguments, verifying citations, or preparing camera-ready submissions. Includes LaTeX templates, reviewer guidelines, and citation verification workflows.
ml-training-recipes
Battle-tested PyTorch training recipes for all domains — LLMs, vision, diffusion, medical imaging, protein/drug discovery, spatial omics, genomics. Covers training loops, optimizer selection (AdamW, Muon), LR scheduling, mixed precision, debugging, and systematic experimentation. Use when training or fine-tuning neural networks, debugging loss spikes or OOM, choosing architectures, or optimizing GPU throughput.
mlflow
Track ML experiments, manage model registry with versioning, deploy models to production, and reproduce experiments with MLflow - framework-agnostic ML lifecycle platform
modal-serverless-gpu
Serverless GPU cloud platform for running ML workloads. Use when you need on-demand GPU access without infrastructure management, deploying ML models as APIs, or running batch jobs with automatic scaling.
model-merging
Merge multiple fine-tuned models using mergekit to combine capabilities without retraining. Use when creating specialized models by blending domain-specific expertise (math + coding + chat), improving performance beyond single models, or experimenting rapidly with model variants. Covers SLERP, TIES-Merging, DARE, Task Arithmetic, linear merging, and production deployment strategies.
model-pruning
Reduce LLM size and accelerate inference using pruning techniques like Wanda and SparseGPT. Use when compressing models without retraining, achieving 50% sparsity with minimal accuracy loss, or enabling faster inference on hardware accelerators. Covers unstructured pruning, structured pruning, N:M sparsity, magnitude pruning, and one-shot methods.
moe-training
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.
nanogpt
Educational GPT implementation in ~300 lines. Reproduces GPT-2 (124M) on OpenWebText. Clean, hackable code for learning transformers. By Andrej Karpathy. Perfect for understanding GPT architecture from scratch. Train on Shakespeare (CPU) or OpenWebText (multi-GPU).
nemo-curator
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.
nemo-evaluator-sdk
Evaluates LLMs across 100+ benchmarks from 18+ harnesses (MMLU, HumanEval, GSM8K, safety, VLM) with multi-backend execution. Use when needing scalable evaluation on local Docker, Slurm HPC, or cloud platforms. NVIDIA's enterprise-grade platform with container-first architecture for reproducible benchmarking.
nemo-guardrails
NVIDIA's runtime safety framework for LLM applications. Features jailbreak detection, input/output validation, fact-checking, hallucination detection, PII filtering, toxicity detection. Uses Colang 2.0 DSL for programmable rails. Production-ready, runs on T4 GPU.
nnsight-remote-interpretability
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.
openrlhf-training
High-performance RLHF framework with Ray+vLLM acceleration. Use for PPO, GRPO, RLOO, DPO training of large models (7B-70B+). Built on Ray, vLLM, ZeRO-3. 2× faster than DeepSpeedChat with distributed architecture and GPU resource sharing.
optimizing-attention-flash
Optimizes transformer attention with Flash Attention for 2-4x speedup and 10-20x memory reduction. Use when training/running transformers with long sequences (>512 tokens), encountering GPU memory issues with attention, or need faster inference. Supports PyTorch native SDPA, flash-attn library, H100 FP8, and sliding window attention.
outlines
Guarantee valid JSON/XML/code structure during generation, use Pydantic models for type-safe outputs, support local models (Transformers, vLLM), and maximize inference speed with Outlines - dottxt.ai's structured generation library
peft-fine-tuning
Parameter-efficient fine-tuning for LLMs using LoRA, QLoRA, and 25+ methods. Use when fine-tuning large models (7B-70B) with limited GPU memory, when you need to train <1% of parameters with minimal accuracy loss, or for multi-adapter serving. HuggingFace's official library integrated with transformers ecosystem.
phoenix-observability
Open-source AI observability platform for LLM tracing, evaluation, and monitoring. Use when debugging LLM applications with detailed traces, running evaluations on datasets, or monitoring production AI systems with real-time insights.
pinecone
Managed vector database for production AI applications. Fully managed, auto-scaling, with hybrid search (dense + sparse), metadata filtering, and namespaces. Low latency (<100ms p95). Use for production RAG, recommendation systems, or semantic search at scale. Best for serverless, managed infrastructure.
presenting-conference-talks
Generates conference presentation slides (Beamer LaTeX PDF and editable PPTX) from a compiled paper with speaker notes and talk script. Use when preparing oral talks, spotlight presentations, or invited talks for ML and systems conferences.
prompt-guard
Meta's 86M prompt injection and jailbreak detector. Filters malicious prompts and third-party data for LLM apps. 99%+ TPR, <1% FPR. Fast (<2ms GPU). Multilingual (8 languages). Deploy with HuggingFace or batch processing for RAG security.
pytorch-fsdp2
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.
pytorch-lightning
High-level PyTorch framework with Trainer class, automatic distributed training (DDP/FSDP/DeepSpeed), callbacks system, and minimal boilerplate. Scales from laptop to supercomputer with same code. Use when you want clean training loops with built-in best practices.
pyvene-interventions
Provides guidance for performing causal interventions on PyTorch models using pyvene's declarative intervention framework. Use when conducting causal tracing, activation patching, interchange intervention training, or testing causal hypotheses about model behavior.
qdrant-vector-search
High-performance vector similarity search engine for RAG and semantic search. Use when building production RAG systems requiring fast nearest neighbor search, hybrid search with filtering, or scalable vector storage with Rust-powered performance.
quantizing-models-bitsandbytes
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.
ray-data
Scalable data processing for ML workloads. Streaming execution across CPU/GPU, supports Parquet/CSV/JSON/images. Integrates with Ray Train, PyTorch, TensorFlow. Scales from single machine to 100s of nodes. Use for batch inference, data preprocessing, multi-modal data loading, or distributed ETL pipelines.
ray-train
Distributed training orchestration across clusters. Scales PyTorch/TensorFlow/HuggingFace from laptop to 1000s of nodes. Built-in hyperparameter tuning with Ray Tune, fault tolerance, elastic scaling. Use when training massive models across multiple machines or running distributed hyperparameter sweeps.
rwkv-architecture
RNN+Transformer hybrid with O(n) inference. Linear time, infinite context, no KV cache. Train like GPT (parallel), infer like RNN (sequential). Linux Foundation AI project. Production at Windows, Office, NeMo. RWKV-7 (March 2025). Models up to 14B parameters.
segment-anything-model
Foundation model for image segmentation with zero-shot transfer. Use when you need to segment any object in images using points, boxes, or masks as prompts, or automatically generate all object masks in an image.
sentence-transformers
Framework for state-of-the-art sentence, text, and image embeddings. Provides 5000+ pre-trained models for semantic similarity, clustering, and retrieval. Supports multilingual, domain-specific, and multimodal models. Use for generating embeddings for RAG, semantic search, or similarity tasks. Best for production embedding generation.
sentencepiece
Language-independent tokenizer treating text as raw Unicode. Supports BPE and Unigram algorithms. Fast (50k sentences/sec), lightweight (6MB memory), deterministic vocabulary. Used by T5, ALBERT, XLNet, mBART. Train on raw text without pre-tokenization. Use when you need multilingual support, CJK languages, or reproducible tokenization.
serving-llms-vllm
Serves LLMs with high throughput using vLLM's PagedAttention and continuous batching. Use when deploying production LLM APIs, optimizing inference latency/throughput, or serving models with limited GPU memory. Supports OpenAI-compatible endpoints, quantization (GPTQ/AWQ/FP8), and tensor parallelism.
sglang
Fast structured generation and serving for LLMs with RadixAttention prefix caching. Use for JSON/regex outputs, constrained decoding, agentic workflows with tool calls, or when you need 5× faster inference than vLLM with prefix sharing. Powers 300,000+ GPUs at xAI, AMD, NVIDIA, and LinkedIn.
simpo-training
Simple Preference Optimization for LLM alignment. Reference-free alternative to DPO with better performance (+6.4 points on AlpacaEval 2.0). No reference model needed, more efficient than DPO. Use for preference alignment when want simpler, faster training than DPO/PPO.
skypilot-multi-cloud-orchestration
Multi-cloud orchestration for ML workloads with automatic cost optimization. Use when you need to run training or batch jobs across multiple clouds, leverage spot instances with auto-recovery, or optimize GPU costs across providers.
slime-rl-training
Provides guidance for LLM post-training with RL using slime, a Megatron+SGLang framework. Use when training GLM models, implementing custom data generation workflows, or needing tight Megatron-LM integration for RL scaling.
sparse-autoencoder-training
Provides guidance for training and analyzing Sparse Autoencoders (SAEs) using SAELens to decompose neural network activations into interpretable features. Use when discovering interpretable features, analyzing superposition, or studying monosemantic representations in language models.
speculative-decoding
Accelerate LLM inference using speculative decoding, Medusa multiple heads, and lookahead decoding techniques. Use when optimizing inference speed (1.5-3.6× speedup), reducing latency for real-time applications, or deploying models with limited compute. Covers draft models, tree-based attention, Jacobi iteration, parallel token generation, and production deployment strategies.
stable-diffusion-image-generation
State-of-the-art text-to-image generation with Stable Diffusion models via HuggingFace Diffusers. Use when generating images from text prompts, performing image-to-image translation, inpainting, or building custom diffusion pipelines.
tensorboard
Visualize training metrics, debug models with histograms, compare experiments, visualize model graphs, and profile performance with TensorBoard - Google's ML visualization toolkit
tensorrt-llm
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.
torchforge-rl-training
Provides guidance for PyTorch-native agentic RL using torchforge, Meta's library separating infra from algorithms. Use when you want clean RL abstractions, easy algorithm experimentation, or scalable training with Monarch and TorchTitan.
training-llms-megatron
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.
transformer-lens-interpretability
Provides guidance for mechanistic interpretability research using TransformerLens to inspect and manipulate transformer internals via HookPoints and activation caching. Use when reverse-engineering model algorithms, studying attention patterns, or performing activation patching experiments.
verl-rl-training
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.
weights-and-biases
Track ML experiments with automatic logging, visualize training in real-time, optimize hyperparameters with sweeps, and manage model registry with W&B - collaborative MLOps platform
whisper
OpenAI's general-purpose speech recognition model. Supports 99 languages, transcription, translation to English, and language identification. Six model sizes from tiny (39M params) to large (1550M params). Use for speech-to-text, podcast transcription, or multilingual audio processing. Best for robust, multilingual ASR.
writing-systems-papers
Paragraph-level structural blueprint for 12-page systems papers targeting OSDI, SOSP, ASPLOS, NSDI, and EuroSys. Provides page allocation, paragraph templates, and writing patterns backed by authoritative guides and best-paper analysis. Complements ml-paper-writing with fine-grained systems-specific guidance.
huggingface-classifier
Hugging Face transformer model fine-tuning and inference for intent classification
rag-embedding-generation
Batch embedding generation with caching, rate limiting, and multiple provider support
ai-framework-watch
Weekly competitive-intelligence digest on the AI agent framework space — momentum, releases, breaking changes across a curated watchlist
audiocraft-audio-generation
PyTorch library for audio generation including text-to-music (MusicGen) and text-to-sound (AudioGen). Use when you need to generate music from text descriptions, create sound effects, or perform melody-conditioned music generation.
distributed-llm-pretraining-torchtitan
Provides PyTorch-native distributed LLM pretraining using torchtitan with 4D parallelism (FSDP2, TP, PP, CP). Use when pretraining Llama 3.1, DeepSeek V3, or custom models at scale from 8 to 512+ GPUs with Float8, torch.compile, and distributed checkpointing.
evaluating-llms-harness
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.
fine-tuning-with-trl
Fine-tune LLMs using reinforcement learning with TRL - SFT for instruction tuning, DPO for preference alignment, PPO/GRPO for reward optimization, and reward model training. Use when need RLHF, align model with preferences, or train from human feedback. Works with HuggingFace Transformers.
gguf-quantization
GGUF format and llama.cpp quantization for efficient CPU/GPU inference. Use when deploying models on consumer hardware, Apple Silicon, or when needing flexible quantization from 2-8 bit without GPU requirements.
guidance
Control LLM output with regex and grammars, guarantee valid JSON/XML/code generation, enforce structured formats, and build multi-step workflows with Guidance - Microsoft Research's constrained generation framework
huggingface-accelerate
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.
huggingface-hub
Hugging Face Hub CLI (hf) — search, download, and upload models and datasets, manage repos, query datasets with SQL, deploy inference endpoints, manage Spaces and buckets.
huggingface-tokenizers
Fast tokenizers optimized for research and production. Rust-based implementation tokenizes 1GB in <20 seconds. Supports BPE, WordPiece, and Unigram algorithms. Train custom vocabularies, track alignments, handle padding/truncation. Integrates seamlessly with transformers. Use when you need high-performance tokenization or custom tokenizer training.
llama-cpp
Runs LLM inference on CPU, Apple Silicon, and consumer GPUs without NVIDIA hardware. Use for edge deployment, M1/M2/M3 Macs, AMD/Intel GPUs, or when CUDA is unavailable. Supports GGUF quantization (1.5-8 bit) for reduced memory and 4-10× speedup vs PyTorch on CPU.
outlines
Outlines: structured JSON/regex/Pydantic LLM generation.
peft-fine-tuning
Parameter-efficient fine-tuning for LLMs using LoRA, QLoRA, and 25+ methods. Use when fine-tuning large models (7B-70B) with limited GPU memory, when you need to train <1% of parameters with minimal accuracy loss, or for multi-adapter serving. HuggingFace's official library integrated with transformers ecosystem.
segment-anything-model
Foundation model for image segmentation with zero-shot transfer. Use when you need to segment any object in images using points, boxes, or masks as prompts, or automatically generate all object masks in an image.
simpo-training
Simple Preference Optimization for LLM alignment. Reference-free alternative to DPO with better performance (+6.4 points on AlpacaEval 2.0). No reference model needed, more efficient than DPO. Use for preference alignment when want simpler, faster training than DPO/PPO.
stable-diffusion-image-generation
State-of-the-art text-to-image generation with Stable Diffusion models via HuggingFace Diffusers. Use when generating images from text prompts, performing image-to-image translation, inpainting, or building custom diffusion pipelines.
transformers
This skill should be used when working with pre-trained transformer models for natural language processing, computer vision, audio, or multimodal tasks. Use for text generation, classification, question answering, translation, summarization, image classification, object detection, speech recognition, and fine-tuning models on custom datasets.
9router-stt
Speech-to-text via 9Router /v1/audio/transcriptions using OpenAI Whisper / Groq / Gemini / Deepgram / AssemblyAI / NVIDIA / HuggingFace models. Use when the user wants to transcribe audio, convert speech to text, or get subtitles from audio files.
add-model-descriptions
Add descriptions for new models from the HuggingFace router to chat-ui configuration. Use when new models are released on the router and need descriptions added to prod.yaml and dev.yaml. Triggers on requests like "add new model descriptions", "update models from router", "sync models", or when explicitly invoking /add-model-descriptions.
v3-core-implementation
Core module implementation for claude-flow v3. Implements DDD domains, clean architecture patterns, dependency injection, and modular TypeScript codebase with comprehensive testing.
unsloth
Expert guidance for fast fine-tuning with Unsloth - 2-5x faster training, 50-80% less memory, LoRA/QLoRA optimization
transformers
This skill should be used when working with pre-trained transformer models for natural language processing, computer vision, audio, or multimodal tasks. Use for text generation, classification, question answering, translation, summarization, image classification, object detection, speech recognition, and fine-tuning models on custom datasets.
00-andruia-consultant
Arquitecto de Soluciones Principal y Consultor Tecnológico de Andru.ia. Diagnostica y traza la hoja de ruta óptima para proyectos de IA en español.
007
Security audit, hardening, threat modeling (STRIDE/PASTA), Red/Blue Team, OWASP checks, code review, incident response, and infrastructure security for any project.
10-andruia-skill-smith
Ingeniero de Sistemas de Andru.ia. Diseña, redacta y despliega nuevas habilidades (skills) dentro del repositorio siguiendo el Estándar de Diamante.
20-andruia-niche-intelligence
Estratega de Inteligencia de Dominio de Andru.ia. Analiza el nicho específico de un proyecto para inyectar conocimientos, regulaciones y estándares únicos del sector. Actívalo tras definir el nicho.
3d-web-experience
Expert in building 3D experiences for the web - Three.js, React Three Fiber, Spline, WebGL, and interactive 3D scenes. Covers product configurators, 3D portfolios, immersive websites, and bringing depth to web experiences.
ab-test-setup
Structured guide for setting up A/B tests with mandatory gates for hypothesis, metrics, and execution readiness.
acceptance-orchestrator
Use when a coding task should be driven end-to-end from issue intake through implementation, review, deployment, and acceptance verification with minimal human re-intervention.
active-directory-attacks
Provide comprehensive techniques for attacking Microsoft Active Directory environments. Covers reconnaissance, credential harvesting, Kerberos attacks, lateral movement, privilege escalation, and domain dominance for red team operations and penetration testing.
activecampaign-automation
Automate ActiveCampaign tasks via Rube MCP (Composio): manage contacts, tags, list subscriptions, automation enrollment, and tasks. Always search tools first for current schemas.
ad-creative
Create, iterate, and scale paid ad creative for Google Ads, Meta, LinkedIn, TikTok, and similar platforms. Use when generating headlines, descriptions, primary text, or large sets of ad variations for testing and performance optimization.
adhx
Fetch any X/Twitter post as clean LLM-friendly JSON. Converts x.com, twitter.com, or adhx.com links into structured data with full article content, author info, and engagement metrics. No scraping or browser required.
advanced-evaluation
This skill should be used when the user asks to "implement LLM-as-judge", "compare model outputs", "create evaluation rubrics", "mitigate evaluation bias", or mentions direct scoring, pairwise comparison, position bias, evaluation pipelines, or automated quality assessment.
advogado-criminal
Advogado criminalista especializado em Maria da Penha, violencia domestica, feminicidio, direito penal brasileiro, medidas protetivas, inquerito policial e acao penal.
advogado-especialista
Advogado especialista em todas as areas do Direito brasileiro: familia, criminal, trabalhista, tributario, consumidor, imobiliario, empresarial, civil e constitucional.
aegisops-ai
Autonomous DevSecOps & FinOps Guardrails. Orchestrates Gemini 3 Flash to audit Linux Kernel patches, Terraform cost drifts, and K8s compliance.
agent-evaluation
Testing and benchmarking LLM agents including behavioral testing, capability assessment, reliability metrics, and production monitoring—where even top agents achieve less than 50% on real-world benchmarks
agent-framework-azure-ai-py
Build persistent agents on Azure AI Foundry using the Microsoft Agent Framework Python SDK.
agent-memory-mcp
A hybrid memory system that provides persistent, searchable knowledge management for AI agents (Architecture, Patterns, Decisions).
agent-memory-systems
Memory is the cornerstone of intelligent agents. Without it, every interaction starts from zero. This skill covers the architecture of agent memory: short-term (context window), long-term (vector stores), and the cognitive architectures that organize them.
agent-orchestration-improve-agent
Systematic improvement of existing agents through performance analysis, prompt engineering, and continuous iteration.
agent-orchestration-multi-agent-optimize
Optimize multi-agent systems with coordinated profiling, workload distribution, and cost-aware orchestration. Use when improving agent performance, throughput, or reliability.
agent-orchestrator
Meta-skill que orquestra todos os agentes do ecossistema. Scan automatico de skills, match por capacidades, coordenacao de workflows multi-skill e registry management.
agent-tool-builder
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.
agentflow
Orchestrate autonomous AI development pipelines through your Kanban board (Asana, GitHub Projects, Linear). Manages multi-worker Claude Code dispatch, deterministic quality gates, adversarial review, per-task cost tracking, and crash-proof pipeline execution.
agentfolio
Skill for discovering and researching autonomous AI agents, tools, and ecosystems using the AgentFolio directory.
agentic-actions-auditor
Audits GitHub Actions workflows for security vulnerabilities in AI agent integrations including Claude Code Action, Gemini CLI, OpenAI Codex, and GitHub AI Inference. Detects attack vectors where attacker-controlled input reaches. AI agents running in CI/CD pipelines.
agentphone
Build AI phone agents with AgentPhone API. Use when the user wants to make phone calls, send/receive SMS, manage phone numbers, create voice agents, set up webhooks, or check usage — anything related to telephony, phone numbers, or voice AI.
docx-official
A user may ask you to create, edit, or analyze the contents of a .docx file. A .docx file is essentially a ZIP archive containing XML files and other resources that you can read or edit. You have different tools and workflows available for different tasks.
accessibility-compliance-accessibility-audit
You are an accessibility expert specializing in WCAG compliance, inclusive design, and assistive technology compatibility. Conduct audits, identify barriers, and provide remediation guidance.
address-github-comments
Use when you need to address review or issue comments on an open GitHub Pull Request using the gh CLI.
agent-manager-skill
Manage multiple local CLI agents via tmux sessions (start/stop/monitor/assign) with cron-friendly scheduling.
avalonia-zafiro-development
Mandatory skills, conventions, and behavioral rules for Avalonia UI development using the Zafiro toolkit.
agentmail
Email infrastructure for AI agents. Create accounts, send/receive emails, manage webhooks, and check karma balance via the AgentMail API.
agents-md
This skill should be used when the user asks to "create AGENTS.md", "update AGENTS.md", "maintain agent docs", "set up CLAUDE.md", or needs to keep agent instructions concise. Enforces research-backed best practices for minimal, high-signal agent documentation.
algorithmic-art
Creating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems. Create original algorithmic art rather than copying existing artists' work to avoid copyright violations.
canvas-design
Create beautiful visual art in .png and .pdf documents using design philosophy. You should use this skill when the user asks to create a poster, piece of art, design, or other static piece. Create original visual designs, never copying existing artists' work to avoid copyright violations.
claude-api
Build, debug, and optimize Claude API / Anthropic SDK apps. Apps built with this skill should include prompt caching. Also handles migrating existing Claude API code between Claude model versions (4.5 → 4.6, 4.6 → 4.7, retired-model replacements). TRIGGER when: code imports `anthropic`/`@anthropic-ai/sdk`; user asks for the Claude API, Anthropic SDK, or Managed Agents; user adds/modifies/tunes a Claude feature (caching, thinking, compaction, tool use, batch, files, citations, memory) or model (Opus/Sonnet/Haiku) in a file; questions about prompt caching / cache hit rate in an Anthropic SDK project. SKIP: file imports `openai`/other-provider SDK, filename like `*-openai.py`/`*-generic.py`, provider-neutral code, general programming/ML.
docx
Use this skill whenever the user wants to create, read, edit, or manipulate Word documents (.docx files). Triggers include: any mention of 'Word doc', 'word document', '.docx', or requests to produce professional documents with formatting like tables of contents, headings, page numbers, or letterheads. Also use when extracting or reorganizing content from .docx files, inserting or replacing images in documents, performing find-and-replace in Word files, working with tracked changes or comments, or converting content into a polished Word document. If the user asks for a 'report', 'memo', 'letter', 'template', or similar deliverable as a Word or .docx file, use this skill. Do NOT use for PDFs, spreadsheets, Google Docs, or general coding tasks unrelated to document generation.
frontend-design
Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, artifacts, posters, or applications (examples include websites, landing pages, dashboards, React components, HTML/CSS layouts, or when styling/beautifying any web UI). Generates creative, polished code and UI design that avoids generic AI aesthetics.
mcp-builder
Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).
Use this skill whenever the user wants to do anything with PDF files. This includes reading or extracting text/tables from PDFs, combining or merging multiple PDFs into one, splitting PDFs apart, rotating pages, adding watermarks, creating new PDFs, filling PDF forms, encrypting/decrypting PDFs, extracting images, and OCR on scanned PDFs to make them searchable. If the user mentions a .pdf file or asks to produce one, use this skill.
pptx
Use this skill any time a .pptx file is involved in any way — as input, output, or both. This includes: creating slide decks, pitch decks, or presentations; reading, parsing, or extracting text from any .pptx file (even if the extracted content will be used elsewhere, like in an email or summary); editing, modifying, or updating existing presentations; combining or splitting slide files; working with templates, layouts, speaker notes, or comments. Trigger whenever the user mentions "deck," "slides," "presentation," or references a .pptx filename, regardless of what they plan to do with the content afterward. If a .pptx file needs to be opened, created, or touched, use this skill.
slack-gif-creator
Knowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack."
theme-factory
Toolkit for styling artifacts with a theme. These artifacts can be slides, docs, reportings, HTML landing pages, etc. There are 10 pre-set themes with colors/fonts that you can apply to any artifact that has been creating, or can generate a new theme on-the-fly.
web-artifacts-builder
Suite of tools for creating elaborate, multi-component claude.ai HTML artifacts using modern frontend web technologies (React, Tailwind CSS, shadcn/ui). Use for complex artifacts requiring state management, routing, or shadcn/ui components - not for simple single-file HTML/JSX artifacts.
webapp-testing
Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs.
xlsx
Use this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved.
hugging-face-jobs
This skill should be used when users want to run any workload on Hugging Face Jobs infrastructure. Covers UV scripts, Docker-based jobs, hardware selection, cost estimation, authentication with tokens, secrets management, timeout configuration, and result persistence. Designed for general-purpose compute workloads including data processing, inference, experiments, batch jobs, and any Python-based tasks. Should be invoked for tasks involving cloud compute, GPU workloads, or when users mention running jobs on Hugging Face infrastructure without local setup.
brand-guidelines
Applies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply.
internal-comms
A set of resources to help me write all kinds of internal communications, using the formats that my company likes to use. Claude should use this skill whenever asked to write some sort of internal communications (status reports, leadership updates, 3P updates, company newsletters, FAQs, incident reports, project updates, etc.).
doc-coauthoring
Guide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks.
security-ownership-map
Analyze git repositories to build a security ownership topology (people-to-file), compute bus factor and sensitive-code ownership, and export CSV/JSON for graph databases and visualization. Trigger only when the user explicitly wants a security-oriented ownership or bus-factor analysis grounded in git history (for example: orphaned sensitive code, security maintainers, CODEOWNERS reality checks for risk, sensitive hotspots, or ownership clusters). Do not trigger for general maintainer lists or non-security ownership questions.
skill-creator
Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.
terradev-gpu-cloud
Cross-cloud GPU provisioning, K8s cluster creation, and inference overflow. Get real-time pricing across 11+ cloud providers, provision the cheapest GPUs in seconds, spin up production K8s clusters, and burst to cloud when your local GPU maxes out. BYOAPI — your keys never leave your machine.
hugging-face-cli
Execute Hugging Face Hub operations using the `hf` CLI. Use when the user needs to download models/datasets/spaces, upload files to Hub repositories, create repos, manage local cache, or run compute jobs on HF infrastructure. Covers authentication, file transfers, repository creation, cache operations, and cloud compute.
ml-engineer
Build production ML systems with PyTorch 2.x, TensorFlow, and modern ML frameworks. Implements model serving, feature engineering, A/B testing, and monitoring. Use PROACTIVELY for ML model deployment, inference optimization, or production ML infrastructure.
evaluating-code-models
Evaluates code generation models across HumanEval, MBPP, MultiPL-E, and 15+ benchmarks with pass@k metrics. Use when benchmarking code models, comparing coding abilities, testing multi-language support, or measuring code generation quality. Industry standard from BigCode Project used by HuggingFace leaderboards.
evaluating-llms-harness
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.
audiocraft-audio-generation
AudioCraft: MusicGen text-to-music, AudioGen text-to-sound.
fine-tuning-expert
Use when fine-tuning LLMs, training custom models, or adapting foundation models for specific tasks. Invoke for configuring LoRA/QLoRA adapters, preparing JSONL training datasets, setting hyperparameters for fine-tuning runs, adapter training, transfer learning, finetuning with Hugging Face PEFT, OpenAI fine-tuning, instruction tuning, RLHF, DPO, or quantizing and deploying fine-tuned models. Trigger terms include: LoRA, QLoRA, PEFT, finetuning, fine-tuning, adapter tuning, LLM training, model training, custom model.
reachy-mini-sdk
Programming guide for Reachy Mini robot using Python SDK v1.8.0 and REST API. Use when controlling Reachy Mini robots, programming movements (head/antennas/body), accessing sensors (camera/microphone/IMU), recording motions, building AI applications, deploying to Hugging Face, or using the daemon REST API. Covers SDK patterns, coordinate systems, interpolation methods, app management, and OpenAPI client generation.
autocli
Use autocli CLI to interact with social/content websites (HackerNews, DevTo, Lobsters, StackOverflow, Steam, Linux-do, Arxiv, Wikipedia, Apple-Podcasts, Xiaoyuzhou, BBC, Hugging Face, SinaFinance, Google, V2EX, Bloomberg, Twitter/X, Bilibili, Reddit, Zhihu, Xiaohongshu, Xueqiu, Weibo, Douban, WeRead, YouTube, Medium, Substack, SinaBlog, BOSS直聘, Jike, Facebook, Instagram, TikTok, Yollomi, Yahoo-Finance, Barchart, LinkedIn, Reuters, SMZDM, Ctrip, Coupang, Grok, Jimeng, Chaoxing, Weixin, Doubao, Cursor, Codex, ChatWise, ChatGPT, Doubao-App, Notion, Discord, Antigravity etc.) via the user's Chrome login session. ALWAYS prefer autocli over playwright/browser automation for these supported sites. Triggers: user asks to browse, search, or fetch hot/trending content from internet, post, or read messages on any web site;
forecasting-reverso
Zero-shot univariate time series forecasting using the Reverso foundation model (NumPy/Numba CPU-only inference). Activate when users provide time series data and request forecasts, predictions, or extrapolations. Supports Reverso Small (550K params). Triggers on "forecast", "predict", "time series", "Reverso", or when tabular data with a temporal dimension needs future-value estimation.
seo-llms-txt
Generate, validate, or audit llms.txt files for AI search visibility. Crawls site structure, generates spec-compliant Markdown index for LLMs. Use when user says "llms.txt", "llm txt", "AI crawlers", "generate llms", "LLM file", "AI readability file".
transformers
This skill should be used when working with pre-trained transformer models for natural language processing, computer vision, audio, or multimodal tasks. Use for text generation, classification, question answering, translation, summarization, image classification, object detection, speech recognition, and fine-tuning models on custom datasets.
academic-aio
Medical AI paper optimization for AI search engines (Perplexity, ChatGPT web, Elicit, Consensus, SciSpace) and RAG-based literature tools. Applies when drafting or reviewing titles, abstracts, structured summary boxes (Key Points / Research in Context / Plain-Language Summary), manuscripts for high-impact medical AI journals (Lancet Digital Health, Radiology, Radiology-AI, npj Digital Medicine, Nature Medicine), preprints (medRxiv/arXiv), GitHub README + CITATION.cff + Zenodo archives, and Hugging Face model/dataset cards. Integrates TRIPOD+AI, CLAIM 2024, STARD-AI, TRIPOD-LLM, DECIDE-AI reporting requirements with generative engine optimization (GEO) principles. Produces a visible pass/fail checklist.
jinja-expert
Author, read, and debug Jinja2 templates across the three places Jinja lives in 2026 — HuggingFace `chat_template.jinja` (rendered by `apply_chat_template` for vLLM / sglang), Ansible playbooks + `.j2` files, and Jinja-adjacent Kubernetes workflows (`values.yaml.j2`, `kubernetes.core.k8s + template`, Helm post-renderers). Companion to the `helm` skill — Helm charts are Go `text/template` + Sprig, not Jinja, and this skill makes that disambiguation explicit.
open-webui-embeddings
Wire HuggingFace embedding + reranker models (BGE-M3, BGE-Reranker-v2-m3, etc.) into Open WebUI's RAG pipeline via LiteLLM proxying HuggingFace Text Embeddings Inference (TEI). Covers the exact wire shapes Open WebUI sends (URL auto-append on embed but NOT rerank; payload + response shapes for both modes), the LiteLLM-TEI gotchas (encoding_format=null trap, HF-driver task_type misdetection, openai vs huggingface driver tradeoffs), TEI config cliffs (max-client-batch-size 422 under hybrid search, max-batch-tokens AS the auto-truncate boundary, arch-specific Docker images), and the end-to-end production config. BGE-M3 + BGE-Reranker-v2-m3 are worked examples; patterns generalise to any TEI encoder.
sglang-model-gateway
SGLang Model Gateway (`sgl-model-gateway`, formerly `sgl-router`) — Rust router fronting vLLM and SGLang inference workers on Kubernetes. Covers first-class vLLM gRPC backend plus HTTP transparent-proxy for vanilla vLLM, the policy set (six `--policy` values, `cache_aware` default), tokenizer-format dispatch (`tokenizer.json` HF-fast vs `tiktoken.model` BPE — including when neither is required because `cache_aware` is text-based), air-gapped recipe (gateway ignores `HF_ENDPOINT`, mount tokenizer files on PVC only when actually needed), K8s manifests with `model_id` labels and per-model RBAC, three HA mitigations (single + PDB, `sessionAffinity: ClientIP`, `--enable-mesh` CRDT sync), and a pitfall catalog covering the Dec 2025 `sgl-router` → `sgl-model-gateway` rename and over-engineered tokenizer init-container traps.
transformers-config-tokenizers-expert
Preflight reference for HuggingFace snapshots — what vLLM, sglang, and transformers.generate see at runtime. Covers config-file precedence (tokenizer.json, tokenizer_config.json, generation_config.json, chat_template.jinja), transformers v5 tokenizer-class taxonomy (TokenizersBackend, PythonBackend, MistralCommonBackend, TikTokenTokenizer), special-token discovery (all_special_ids, added_tokens_decoder, extra_special_tokens, backend_tokenizer.get_added_tokens_decoder), chat-template Jinja contract (ImmutableSandboxedEnvironment, loopcontrols, raise_exception, strftime_now, tojson, add_generation_prompt), and engine knobs (skip_special_tokens, trust_request_chat_template, chat_template_kwargs allowlist, adjust_request, incremental detokenizer, EOS merge). Ships verified 2026 hall-of-shame for Kimi-K2.6, GLM-5.1, Gemma-4, Qwen3, DeepSeek-V3, plus drop-in Python for resolving markers to IDs, detecting turn-primer-as-EOS leaks, and cross-referencing tokenizer.json vs tokenizer_config.json.
vllm-configuration
Configure vLLM completely — YAML config file format, CLI arg precedence, full VLLM_*/HF_*/TRANSFORMERS_* env-var catalog, end-to-end recipe for air-gapped environments (internal HF mirrors, hf-mirror.com, ModelScope, HF_HUB_OFFLINE with pre-seeded cache, gated models offline, trust_remote_code supply-chain implications). VLLM_HOST_IP vs API-host confusion, Kubernetes-service-named-`vllm` env-var poisoning, usage-stats triple opt-out, YAML precedence surprises.
hugging-face-jobs
This skill should be used when users want to run any workload on Hugging Face Jobs infrastructure. Covers UV scripts, Docker-based jobs, hardware selection, cost estimation, authentication with tokens, secrets management, timeout configuration, and result persistence. Designed for general-purpose compute workloads including data processing, inference, experiments, batch jobs, and any Python-based tasks. Should be invoked for tasks involving cloud compute, GPU workloads, or when users mention running jobs on Hugging Face infrastructure without local setup.
hugging-face-model-trainer
This skill should be used when users want to train or fine-tune language models using TRL (Transformer Reinforcement Learning) on Hugging Face Jobs infrastructure. Covers SFT, DPO, GRPO and reward modeling training methods, plus GGUF conversion for local deployment. Includes guidance on the TRL Jobs package, UV scripts with PEP 723 format, dataset preparation and validation, hardware selection, cost estimation, Trackio monitoring, Hub authentication, and model persistence. Should be invoked for tasks involving cloud GPU training, GGUF conversion, or when users mention training on Hugging Face Jobs without local GPU setup.
local-rag-builder
本地 RAG 系统搭建技能，支持环境检测修复、嵌入模型多源下载、5种切分策略 + GuardStack + 后处理 + 插件注册、多知识库管理 + 自动分类规则、可调 Prompt、Web 可视化配置 + 极客模式 + 模板管理
audiocraft-audio-generation
AudioCraft: MusicGen text-to-music, AudioGen text-to-sound.
evaluating-llms-harness
lm-eval-harness: benchmark LLMs (MMLU, GSM8K, etc.).
huggingface-hub
HuggingFace hf CLI: search/download/upload models, datasets.
llama-cpp
llama.cpp local GGUF inference + HF Hub model discovery.
segment-anything-model
SAM: zero-shot image segmentation via points, boxes, masks.
ai-tools
Provides guidance for integrating AI tools and components into the Family Tree App, including knowledge graphs, computer vision, and natural language processing. Invoke when working on AI-related features or when needing AI integration advice.
offensive-osint
Operational arsenal for external red-team and bug-bounty reconnaissance. Concrete wordlists (28 Swagger paths, 13 GraphQL paths, 35 high-risk ports, 6 missing-header findings, 15 always-on HTTP checks, 5 SAML paths, cloud bucket permutations, JS guess-paths, vendor product fingerprints for Citrix/F5/Pulse/Fortinet/Cisco/PaloAlto/VMware/Exchange, cloud-native service fingerprints, container/K8s exposure paths, CI/CD platform paths, documentation/wiki leak paths, WHOIS/RDAP, DNS record catalog, Wayback CDX recipes), 43+-pattern secret-regex catalog (incl. modern AI API keys: Anthropic/OpenAI/HuggingFace/Cloudflare/DigitalOcean/npm/PyPI/Docker Hub/Atlassian/DataDog/Sentry/ngrok), 80+ dork corpus across 9 categories, GitHub code-search dorks, copy-paste curl/httpie probes for every check, post-discovery enumeration workflows (AWS/GitHub/Slack/JWT/PMAK/Anthropic/OpenAI), endpoint interest scoring rubric (0–100), mobile app ownership confidence, identity-fabric endpoints (Entra/Okta/ADFS/Google/SAML/M365 Teams+Shar
hugging-face-papers
Read and analyze Hugging Face paper pages or arXiv papers with markdown and papers API metadata.
dependency-extraction-multilang
Extract dependencies from package.json, requirements.txt, and pyproject.toml. Use when analyzing repo tech stacks, building dependency graphs, or matching projects by technology.
github-discovery-scoring
Score GitHub repos for project relevance using weighted dependency/topic matching. Use when building recommendation engines, filtering discovery results, or ranking repositories.
scriba
Transcribe any audio/video meeting into an accurate, speaker-diarized Markdown transcript (Speaker 1/2/3), then rename speakers to real names. Use when the user drops an audio or video file and wants a transcript, расшифровку, транскрипт, диаризацию, "кто что сказал", or asks to транскрибировать/расшифровать встречу или запись.
hf-mcp
Use Hugging Face Hub via MCP server tools. Search models, datasets, Spaces, papers. Get repo details, fetch documentation, run compute jobs, and use Gradio Spaces as AI tools. Available when connected to the HF MCP server.
hugging-face-cli
Execute Hugging Face Hub operations using the `hf` CLI. Use when the user needs to download models/datasets/spaces, upload files to Hub repositories, create repos, manage local cache, or run compute jobs on HF infrastructure. Covers authentication, file transfers, repository creation, cache operations, and cloud compute.
hugging-face-dataset-viewer
Use this skill for Hugging Face Dataset Viewer API workflows that fetch subset/split metadata, paginate rows, search text, apply filters, download parquet URLs, and read size or statistics.
hugging-face-datasets
Create and manage datasets on Hugging Face Hub. Supports initializing repos, defining configs/system prompts, streaming row updates, and SQL-based dataset querying/transformation. Designed to work alongside HF MCP server for comprehensive dataset workflows.
hugging-face-evaluation
Add and manage evaluation results in Hugging Face model cards. Supports extracting eval tables from README content, importing scores from Artificial Analysis API, and running custom model evaluations with vLLM/lighteval. Works with the model-index metadata format.
hugging-face-paper-publisher
Publish and manage research papers on Hugging Face Hub. Supports creating paper pages, linking papers to models/datasets, claiming authorship, and generating professional markdown-based research articles.
hugging-face-tool-builder
Use this skill when the user wants to build tool/scripts or achieve a task where using data from the Hugging Face API would help. This is especially useful when chaining or combining API calls or the task will be repeated/automated. This Skill creates a reusable script to fetch, enrich or process data.
hugging-face-trackio
Track and visualize ML training experiments with Trackio. Use when logging metrics during training (Python API), firing alerts for training diagnostics, or retrieving/analyzing logged metrics (CLI). Supports real-time dashboard visualization, alerts with webhooks, HF Space syncing, and JSON output for automation.
codeclaw
Export Claude Code and Codex conversation history to Hugging Face as structured training data. Use when the user asks about exporting conversations, uploading to Hugging Face, configuring codeclaw, reviewing PII/secrets in exports, or managing their dataset.
deploy-kimi-k26-on-rtx-pro-6000
Deploy and serve Moonshot Kimi-K2.6 (1T MoE, MLA, 256K context, vision) in a user-chosen quantization — official INT4 QAT (moonshotai/Kimi-K2.6, compressed-tensors→Marlin; vLLM or SGLang) or NVFP4 (nvidia/Kimi-K2.6-NVFP4, ModelOpt FP4; vLLM only — SGLang NVFP4 is NaN-broken on sm_120) — on a Linux server (verified Ubuntu 26.04) with 8× NVIDIA RTX PRO 6000 Blackwell Server Edition (96 GB, sm_120) GPUs. The quantization and the engine are both chosen at deploy time with a hardware-based recommendation. Runs an official-image Docker container via nvidia-container-toolkit CDI (--device nvidia.com/gpu=all --ipc=host --network host, bind-mounted weights), exposing an OpenAI-compatible API on :30000 behind one static systemd service `kimi-k26` (quant + engine selected via its EnvironmentFile — only one 595 GB variant fits the 8-GPU pool at a time). Use when deploying or serving Kimi-K2.6 INT4 or NVFP4 on RTX PRO 6000 Blackwell / sm_120 hardware (vLLM-in-Docker, or SGLang-in-Docker for INT4) — or troubleshooting NCCL
nlp-pretraining
Best practices for language model pretraining and fine-tuning. Use when generating or reviewing NLP training code.
agents-md-pro
Create, optimize, update, and validate AGENTS.md files with maximum token efficiency. Use when the user asks to (1) create new AGENTS.md files for any repository, (2) optimize/condense existing AGENTS.md to reduce token count, (3) update/refresh AGENTS.md to sync with codebase changes, (4) validate AGENTS.md quality and completeness, or (5) improve AGENTS.md files to be more effective for AI agents. Always generates token-efficient, condensed output focused on actionable commands and patterns while maintaining model-agnostic language.
troubleshoot
Diagnose common cqs issues — stale index, model download, schema mismatch, connection issues.
packagebuilder
Build conformant arc-installable packages for the metafactory ecosystem. Encodes all conventions from arc manifests, blueprint tracking, compass governance, content-filter safety, test-rig verification, trust metadata, and PR quality standards. USE WHEN build package, create package, new package, author-builder, package skill, metafactory package, arc package, create skill, create tool, create agent, create component, submit package, package review, package conventions, how to build for metafactory.
paper-data
Use this skill when the user wants to find datasets, select datasets for their research, process data, check for data leakage, or prepare data for experiments. Triggers include: "find datasets", "which dataset", "data processing", "data pipeline", "data leakage", "data split", "prepare data". Also use when setting up a data processing pipeline or checking data quality.
paper-plan
Use this skill when the user wants to design experiments, plan an experiment campaign, find open-source code or datasets for their research, or create a project plan. Triggers include: "design experiments", "plan experiments", "experiment design", "how to run experiments", "find code", "find datasets", "project plan". Also use when translating an idea into a concrete, executable experiment plan.
opc
多引擎 AI 创作工具链。TTS 语音合成（edge-tts / Qwen3-TTS）、ASR 语音识别与卡拉OK字幕、ComfyUI 图片生成（ERNIE/Qwen/Z-Image + PromptKG 知识图谱）、视频剪辑（Cut）。使用场景：(1) 文本转语音播放，(2) 音频转录生成 SRT/ASS 字幕，(3) AI 图片生成与风格探索，(4) Prompt 知识图谱查询与模板发现，(5) 字幕级视频剪辑。触发词：语音、TTS、ASR、字幕、图片生成、prompt、知识图谱、KG、视频剪辑、cut
evolve
Evolutionary search for agent improvement using Mind Evolution. Manages multiple parallel improvement islands via git worktrees, with selection and cross-pollination. Invoke with /evolve, "evolve my agent", "run evolutionary search", "parallel improvement".
ratchet
Autonomous ratchet loop for agent improvement. Configures optimization targets, then loops: improve agent → run agent → eval → keep or revert. Uses the /recursive-improve pipeline internally with auto-approval. Invoke with /ratchet or "run the ratchet loop", "improve my agent overnight", "autonomous improvement".
recursive-improve
End-to-end agent improvement pipeline. Analyzes raw execution traces, extracts insights, manages a skillbook, gathers domain context, defines metrics, builds a rubric, creates a prioritized action plan, presents it for review, and implements approved fixes. Trigger when the user says "improve my agent", "run the improvement pipeline", "apply insights", "/recursive-improve", or when eval/traces/ contains trace files.
ai-creative-tools
Orchestrate advanced creative AI workflows using ComfyUI, ElevenLabs, and Hugging Face integrations. Triggers on "generate AI art", "ComfyUI workflow", "ElevenLabs voice", "automate creative pipeline", "generate asset queue", or "AI image generation".
kronos-agent
Financial time-series forecasting using the Kronos foundation model (MIT, NeoQuasar). Takes OHLC candles, returns predicted future candles with configurable horizon. Infrastructure skill — called by trader agents or scheduled ingestion, not directly invoked by users or the model.
paper-env
Use this skill when the user wants to set up a research environment, install dependencies, configure GPU drivers, or configure a new machine for experiments. Triggers include: "setup environment", "install dependencies", "configure GPU", "CUDA setup", "environment config", "server setup", "new machine setup". Also use when encountering environment-related errors or version conflicts.
research-collaborator
Use this skill whenever a researcher wants to test, validate, stress-test, or falsify a research idea or hypothesis — especially in AI/ML/deep learning. Trigger on phrases like "I have an idea," "would this work," "test this hypothesis," "sanity check my idea," "what's wrong with this idea," "review my results," "is this publishable," "why isn't this working," or any request to evaluate the feasibility, novelty, or correctness of a research concept.
benchmark
Run metric quality benchmark, store results, and compare against previous runs. Invoke with /benchmark, "run benchmark", "benchmark metrics", "check metric quality".
gemini-manager
This skill should be used when the user wants Claude Code to act purely as a manager/architect while Gemini CLI does all the coding work. Claude Code drives Gemini like an intern - issuing tasks, reviewing output, requesting fixes - but never writes code itself. Use when user says "manage gemini", "architect mode", "drive gemini", or wants to delegate all implementation to Gemini.
dataclaw-sync
增量同步 AI 对话记录到 Obsidian。使用 DataClaw 导出 Claude Code、OpenCode、Codex、Kimi、Gemini CLI、OpenClaw 等 AI Agent 的对话，转换为 Obsidian 笔记存入 LabNotes vault。支持增量更新，只处理新增会话。可选上传到 Hugging Face。当用户说"同步对话"、"导出对话"、"sync conversations"、"对话存档到 Obsidian"、"/dataclaw-sync" 时触发。
colab-video-pipeline
Use this skill when running or maintaining the Jiang Lens Google Colab video pipeline for YouTube download, diarization, transcription, Drive sync, or Playwright-based Colab automation. Requires the project Drive folder named jianglens and never commits cookies, browser profiles, tokens, or downloaded media.
results-to-slides
Automate the grunt work of making research presentations — discovers experiments from git/output folders, collects images and metrics, and organizes them into slides. Creates a slide-by-slide script for user approval, then generates slide markdown and editable PPTX.
long-context
Extend context windows of transformer models using RoPE, YaRN, ALiBi, and position interpolation techniques. Use when processing long documents (32k-128k+ tokens), extending pre-trained models beyond original context limits, or implementing efficient positional encodings. Covers rotary embeddings, attention biases, interpolation methods, and extrapolation strategies for LLMs.
long-context
Extend context windows of transformer models using RoPE, YaRN, ALiBi, and position interpolation techniques. Use when processing long documents (32k-128k+ tokens), extending pre-trained models beyond original context limits, or implementing efficient positional encodings. Covers rotary embeddings, attention biases, interpolation methods, and extrapolation strategies for LLMs.
long-context
Extend context windows of transformer models using RoPE, YaRN, ALiBi, and position interpolation techniques. Use when processing long documents (32k-128k+ tokens), extending pre-trained models beyond original context limits, or implementing efficient positional encodings. Covers rotary embeddings, attention biases, interpolation methods, and extrapolation strategies for LLMs.
long-context
Extend context windows of transformer models using RoPE, YaRN, ALiBi, and position interpolation techniques. Use when processing long documents (32k-128k+ tokens), extending pre-trained models beyond original context limits, or implementing efficient positional encodings. Covers rotary embeddings, attention biases, interpolation methods, and extrapolation strategies for LLMs.
brainstorming
Collaborative design exploration that refines ideas into validated specs through iterative questioning. Use before any creative work including creating features, building components, adding functionality, or modifying behavior.
dispatching-parallel-agents
Dispatches one subagent per independent domain to parallelize investigation/fixes. Use when you have 2+ unrelated failures (e.g., separate failing test files, subsystems, bugs) with no shared state or ordering dependencies.
executing-plans
Disciplined plan execution for implementation tasks. Use when executing a saved implementation plan, following step-by-step instructions from a plan document.
finishing-a-development-branch
Git branch completion workflow. Use when implementation is complete, tests pass, and a feature branch needs to be integrated via merge, pull request, or cleanup.
receiving-code-review
Assesses and responds to incoming code review feedback on PRs (reviewer comments, requested changes), especially when suggestions are unclear, technically questionable, or scope-expanding. Use before implementing review suggestions to align on intent and keep changes minimal.
requesting-code-review
Use when you need to request a code review for a PR/MR and want a consistent review brief (context, scope, risk areas, test instructions, acceptance criteria) before merge.
subagent-driven-development
Sequential subagent execution with two-stage review gates for implementation plans. Use when executing multi-task plans in current session, when tasks need fresh subagent context to avoid pollution, when formal review cycles (spec compliance then code quality) are required between tasks, or when you need diff-based validation of each task before proceeding.
systematic-debugging
Root cause analysis for debugging. Use when bugs, test failures, or unexpected behavior have non-obvious causes, or after multiple fix attempts have failed.
test-driven-development
Red-green-refactor development methodology requiring verified test coverage. Use for feature implementation, bugfixes, refactoring, or any behavior changes where tests must prove correctness.
using-git-worktrees
Git worktree–based workspace isolation for parallel or non-disruptive development. Use when work must occur without modifying or interfering with the current working tree.
using-superpowers
Meta-skill enforcing skill discovery and invocation discipline through mandatory workflows. Use when starting any conversation to check for relevant skills before any response, ensuring skill-first workflow before proceeding.
verification-before-completion
Verification discipline for completion claims. Use when about to assert success, claim a fix is complete, report tests passing, or before commits and PRs. Enforces evidence-first workflow.
writing-plans
Structured implementation planning for multi-step development tasks. Use when you have a spec or requirements and need to break work into executable steps.
writing-skills
Use when creating new skills, editing existing skills, or verifying skills work before deployment
model-cards
Use when authoring or interpreting Mitchell-extended model cards in this plugin. Covers when each of the 10 sections applies, the citation discipline, the honesty rules (claim-level and card-level), and the tiered source strategy. Reference for the model-card-researcher agent and the /model-card command.
citation-validator
Use this skill when modifying src/regulaitor/citation/validator.py or its policy. Documents the canonical 3-check validation procedure and the rules for evolving it (e.g. adding a fuzzy-fallback layer in H15).
document-analysis
Use this skill when extracting, sanitizing, segmenting, or analyzing a document end-to-end through the RegulAItor pipeline (PDF or Markdown). Activates the full extract→sanitize→segment→loop[gate→retriever→analyst→auditor]→aggregate flow with SSDLC-aligned defaults.
evals-runner
Use this skill when running the H8 evaluation harness, reading `evals/reports/latest.md`, deciding whether to re-run, or extending the gold set. Activates from H8 onwards.
prompt-versioning
Use this skill when adding, modifying, or rolling back agent prompts in src/regulaitor/agents/prompts/ to keep the project's prompt history reproducible and auditable. Activates from H4 onwards (Analyst, Auditor, Council).
rag-ingest
Use this skill when adding a new regulatory corpus (NIS2, DORA, or any future norma) following the H1 RegulAItor pattern. Ensures the new corpus integrates with the existing fetch/parse/validate/manifest pipeline without ad-hoc divergence.
redteam-runner
Use this skill when running the H9 red team suite, reading `redteam/reports/latest.md`, deciding when to re-run, or extending the attack set. Activates from H9 onwards.
secure-coding-checklist
Use this skill before merging any PR that touches src/regulaitor/security/, src/regulaitor/document/sanitizer.py, src/regulaitor/citation/validator.py, or src/regulaitor/agents/auditor.py. Activates from H9 onwards (CLAUDE.md §12.3.10).
fused-integrations
Reference for using Fused's built-in integration connections inside UDFs. Covers data sources (Snowflake, BigQuery, GCS, S3, Airtable, Notion, Google Drive), compute/inference providers (Modal, Hugging Face, Baseten, Daytona, ComfyOrg, Slack), and LLM providers (Anthropic, OpenAI) — the fused.api connect helpers, secrets access, and common operations (query, write, list, invoke, infer). Use when the user is writing a UDF that reads from, writes to, or calls out to a connected service.
nvidia-developer-firehose
Real-time AI-ECOSYSTEM firehose (v3 multi-source, May 2026). Polls 12 Atom/RSS feeds every 30 min: NVIDIA (developer-blog + main-blog + newsroom), hyperscalers (Azure, AWS, AWS-ML, Meta-Engineering), AI labs (OpenAI, DeepMind, Hugging Face), and neoclouds (CoreWeave, Together AI). For each new post, uses HEURISTIC EXTRACTION + yfinance.Search to auto-resolve every mentioned company → US ticker (no hand-maintained name→ticker dict), with a persistent ticker cache that learns over time. Surfaces names via Telegram tagged by source, separated into 🎯 portfolio-tracked tickers vs 🔍 newly discovered tickers. Why it matters: hyperscalers + NVIDIA + AI labs publicly name 800V HVDC, CPO, optical, power, and custom-silicon partners — the forward-looking design ecosystem that re-rates weeks later when sell-side picks it up. (AMD has no public RSS — covered separately by SEC 8-K strategic-partner-firehose.) Triggers in English ("nvidia developer firehose", "ai ecosystem firehose", "ai partner monitor", "hyperscaler blo
aio-remove-background
Remove image backgrounds via RMBG-2.0 alpha matting with despill + 1px erode — produces clean PNG RGBA cutout that handles hair, smoke, glow, soft edges. Use when the user wants to remove background, xoá nền, make transparent, cutout, chroma key, or post-process a text-to-image render (gpt-image, Imagen, FLUX, SDXL, Grok, Midjourney) — especially with flat magenta (#FF00FF) key.
huggingface-dataset-publishing
当需要创建、上传、验证和维护 Hugging Face Dataset 时使用，尤其是包含图片、多图字段、文本步骤、答案、JSON/metadata 字段的数据集；包括 datasets Features/Image/Sequence 的规范用法、push_to_hub、上传后 load_dataset 回读验证、Dataset Viewer 图片浏览检查、clone dataset repo 后用 git 维护 README/dataset card，以及 token 安全。
lab-cluster-1
当需要在 lab cluster 1 / PJLAB 上使用开发机、rlaunch worker 或 rjob 任务时使用；覆盖交互 SSH、安全边界、路径规范、代理、CPU/GPU 分区、训练/部署、服务访问和排错，并要求使用原始 rlaunch/rjob 命令。
model-tampering
AI model supply chain attack methodology covering weight tampering, malicious fine-tuning backdoor insertion, plugin/extension hijacking, and model provenance verification bypass. For authorized assessments of AI deployment pipelines.
autocli
Use autocli CLI to interact with social/content websites (HackerNews, DevTo, Lobsters, StackOverflow, Steam, Linux-do, Arxiv, Wikipedia, Apple-Podcasts, Xiaoyuzhou, BBC, Hugging Face, SinaFinance, Google, V2EX, Bloomberg, Twitter/X, Bilibili, Reddit, Zhihu, Xiaohongshu, Xueqiu, Weibo, Douban, WeRead, YouTube, Medium, Substack, SinaBlog, BOSS直聘, Jike, Facebook, Instagram, TikTok, Yollomi, Yahoo-Finance, Barchart, LinkedIn, Reuters, SMZDM, Ctrip, Coupang, Grok, Jimeng, Chaoxing, Weixin, Doubao, Cursor, Codex, ChatWise, ChatGPT, Doubao-App, Notion, Discord, Antigravity etc.) via the user's Chrome login session. ALWAYS prefer autocli over playwright/browser automation for these supported sites. Triggers: user asks to browse, search, or fetch hot/trending content from internet, post, or read messages on any web site;
Integration detected automatically from skill content. Some results may be false positives.