Skill UI

Sentence Transformers provides 5000+ pre-trained models that generate sentence, text, and image embeddings for RAG, semantic search, clustering, and multilingual production workloads without relying on external APIs.

TransformerLens is the standard library for performing mechanistic interpretability research on large language models. It provides clean interfaces to inspect and manipulate transformer internals—such as attention patterns, residual streams, and MLP outputs—via activation caching and HookPoints. This tool is essential for reverse-engineering model algorithms, performing causal tracing, and studying the internal circuits responsible for model behavior.

Reasoning-Driven Image Briefs skill transforms agent prompts into logic-based Gemini 3 style instructions for muapi.ai, adding structured creative briefs (subject/action/context/composition/lighting), precise text cues, and physics-grounded constraints before invoking generation scripts.

PyHealth delivers a full-stack Python toolkit for building, evaluating, and deploying clinical machine learning models, covering EHR datasets, medical coding systems, physiological signals, and healthcare-tailored architectures such as RETAIN, SafeDrug, and Transformer-based pipelines.

Reference for building Graph Neural Networks with PyTorch Geometric, covering data objects, datasets, transforms, convolution layers, lazy initialization, and heterogeneous support so you can quickly tackle node, link, and graph tasks.

A comprehensive skill for leveraging massive pre-trained transformer models across diverse AI domains, including Natural Language Processing (NLP), Computer Vision (CV), and Audio. Use this skill for advanced tasks such as text generation, classification, question answering, summarization, object detection, and model fine-tuning on proprietary datasets. It provides streamlined pipelines and granular control for deployment.

Transform briefs into cinematic videos via Seedance 2.0, guiding text/image/video/audio inputs through director-level directives, multimodal tagging, and synchronized sound for 4-15s outputs.

Transforms vague or inconsistent prompts into precision-engineered instructions so any LLM delivers structured, hallucination-resistant, token-efficient outputs.

Manage Hugging Face Hub datasets with repo creation, configuration storage, streaming updates, and DuckDB SQL queries/transformations, letting you extend MCP workflows with rich editing capabilities.

Deploys AI and NLP-powered detection to spot business email compromise by analyzing writing style, behavioral context, and intent, then integrates baselines, transformer models, and response workflows to surface impersonation attempts that rule-based filters miss.

Transforms a refined research proposal or method idea into a concise, claim-backed experiment roadmap, covering ablations, evaluation protocols, compute budgets, and execution order so papers can defend the problem, novelty, and any LLM/VLM/Diffusion/RL contribution.

Run Hugging Face models directly in JavaScript/TypeScript for text, vision, and speech tasks without backend servers, using Transformers.js in Node.js or modern browsers with WebGPU, quantization, and pipeline abstractions for efficient inference.