Skill UI

This skill manages the entire academic rebuttal cycle. It parses external reviews from major conferences (like ICML/OpenReview), drafts safe, text-only responses while strictly enforcing grounding (no fabrication) and full issue coverage. It supports multi-round follow-ups and adheres to venue-specific constraints, ensuring every claim is provable.

A zero-config, serverless cloud platform designed for running intensive GPU workloads, including model training, fine-tuning, and inference. It eliminates the need for complex infrastructure setup (no SSH, no Docker) while supporting local-first development and auto-scaling for various high-performance GPUs (A100, H100). Ideal for data scientists and ML engineers requiring scalable and cost-efficient remote compute resources.

Sanitize user prompts through Google Model Armor templates via the gws CLI command, ensuring outbound safety and supporting stdin, text, or explicit JSON inputs for prompt filtering.

A powerful tool leveraging OpenRouter and the Gemini 3 Pro Image model to create, modify, and composite high-quality images. Supports three primary use cases: prompt-only generation from text, detailed single-image edits (inpainting/outpainting), and complex multi-image composition. Users can specify resolutions up to 4K, making it suitable for professional creative and design workflows.

This skill guides users on setting up and running LLM-as-judge evaluations using Arize. It covers defining evaluators (including prompts, classification choices, and models) and executing tasks against real data spans or experiments. It supports key metrics like hallucination, faithfulness, and correctness, and facilitates continuous monitoring for comprehensive LLM performance analysis.

Three-layer verification pipeline that extracts verifiable claims from AI output, searches the web for supporting or contradicting sources, flags hallucinations, and builds structured reports for human review.

Utilize advanced media intelligence to analyze audio, image, video, and text for synthetic manipulation and AI-generated content. This skill detects deepfakes, traces the original AI synthesis source, applies and detects watermarks, verifies speaker identity, and provides comprehensive media provenance analysis, ensuring media authenticity claims are rigorously supported.

DiffDock is a cutting-edge deep learning tool leveraging diffusion models to predict the 3D binding poses of small molecule ligands to protein targets. It supports structure-based drug discovery, enabling single-complex docking, batch virtual screening, and lead optimization. Users can input protein structures (PDB) or sequences, along with diverse ligand formats (SMILES, SDF, MOL2). The tool outputs predicted poses and confidence scores, making it crucial for identifying potential drug candidates, though it does not predict binding affinity.

Open Notebook is a self-hosted, open-source alternative to major AI research tools. It allows users to ingest diverse content—including PDFs, videos, and web pages—and organize them into private notebooks. Key features include context-aware AI chat, automated summarization, multi-speaker podcast generation, and full programmatic access via a comprehensive REST API. By running on your own infrastructure, it ensures complete data privacy and supports 16+ AI providers.

PufferLib is a high-performance, scalable RL framework designed for massive parallel training and vectorized simulation. Featuring PuffeRL (optimized PPO+LSTM) and the PufferEnv API, it supports multi-agent systems, custom environment creation, and achieves industry-leading speeds (up to millions of steps/second). Ideal for advanced research and large-scale AI experimentation.

A comprehensive framework designed to structure and automate PyTorch deep learning workflows. It simplifies boilerplate code by organizing models into LightningModules, automating training loops via the Trainer, and providing robust support for multi-GPU/TPU scaling (DDP, FSDP, DeepSpeed), data handling (DataModules), and best-in-class experiment logging (W&B, MLflow). Ideal for professional, scalable deep learning project development.

A comprehensive Python framework utilizing deep generative models and variational inference for advanced single-cell genomics analysis. It supports the integration and analysis of diverse multi-modal data types, including single-cell RNA-seq, ATAC-seq, CITE-seq, and spatial transcriptomics. Key applications include probabilistic batch correction, cell type annotation, and multi-omics integration.