MLOps
Operacions d’aprenentatge automàtic
data-poisoning
pluginagentmarketplace
Test AI training pipelines for data poisoning vulnerabilities and backdoor injection
did-core
hairyf
W3C DID Core specification—DID/DID URL syntax, data model, core properties, verification methods, services, representations, and DID methods.
benchmark-datasets
pluginagentmarketplace
Standard datasets and benchmarks for evaluating AI security, robustness, and safety
funsloth-upload
chrisvoncsefalvay
Generate comprehensive model cards and upload fine-tuned models to Hugging Face Hub with professional documentation
Healthsafety
robdtaylor
UK health and safety management for manufacturing operations. Covers risk assessments, COSHH, PPE, incident reporting, LOTO, manual handling, and emergency procedures. Aligned with UK HSE regulations and IATF 16949 clause 7.1.4. USE WHEN user says 'health and safety', 'risk assessment', 'COSHH', 'PPE', 'accident', 'incident', 'LOTO', 'lockout tagout', 'manual handling', 'emergency procedure', or needs safety compliance guidance.
funsloth-train
chrisvoncsefalvay
Generate Unsloth training notebooks and scripts. Use when the user wants to create a training notebook, configure fine-tuning parameters, or set up SFT/DPO/GRPO training.
funsloth-hfjobs
chrisvoncsefalvay
Training manager for Hugging Face Jobs - launch fine-tuning on HF cloud GPUs with optional WandB monitoring
model-inversion
pluginagentmarketplace
Privacy attacks to extract training data and sensitive information from AI models
model-extraction
pluginagentmarketplace
Techniques to extract model weights, architecture, and training data through API queries
funsloth-local
chrisvoncsefalvay
Training manager for local GPU training - validate CUDA, manage GPU selection, monitor progress, handle checkpoints
hytale-custom-blocks
MnkyArts
Create custom block types for Hytale with textures, physics, states, farming, and interactions. Use when asked to "add a custom block", "create a new block type", "make blocks farmable", "add block interactions", or "configure block physics".
concept-explorer
Semprini
Use this skill when the user asks "what is [concept]", "how does [feature] work", "explain [section]", "compare MD-DDL to [tool]", "why does MD-DDL [design choice]", or any question about a specific part of the standard. Also use when the user wants to understand the reasoning behind a design decision or the trade-offs between alternatives (entity vs enum, canonical vs bounded context, etc.).
fine-tuning-with-trl
magnetoid
"TRL: SFT, DPO, PPO, GRPO, reward modeling for LLM RLHF."
recent-form-review
tomevault-io
Review a validated recent-match summary and deterministic report using attributable RiftCoach knowledge. Use for recent form, recent match trends, recurring weaknesses, or short training-priority requests. Use when this capability is needed.
stable-diffusion-image-generation
magnetoid
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.
modal-serverless-gpu
magnetoid
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.
huggingface-tokenizers
magnetoid
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.
orientation
Semprini
Use this skill on first contact, when the user asks "what is MD-DDL", "where do I start", "what can I do here", or describes their role and goals. Also use when the user asks for a general overview of the standard, the agent ecosystem, or the workflow.
lambda-labs-gpu-cloud
magnetoid
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.
scvi-tools
Victory-Hugo
单细胞组学深度生成模型。当需要概率批次校正(scVI)、迁移学习、带不确定性的差异表达或多模态集成(TOTALVI, MultiVI)时使用。最适合高级建模、批次效应、多模态数据。标准分析管道请使用scanpy。
tensorrt-llm
magnetoid
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.
nemo-curator
magnetoid
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.
go-create-gorm-model
cristiano-pacheco
Generate Go GORM models following Pingo modular architecture conventions. Use when creating or updating persistence models in internal/modules/<module>/model/, including table mapping, nullable SQL types, timestamps, and relation fields for identity and monitor modules.
distributed-llm-pretraining-torchtitan
magnetoid
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.