Skill UI

Assists with tensorboard visualizer tasks for ML training by guiding setup, best practices, and experiment tracking whenever tensorboard visualization is mentioned, producing production-ready configs and validated outputs.

Optimizes large language model inference on NVIDIA GPUs with TensorRT, delivering 10-100× faster throughput, quantized precision (FP8/INT4), multi-GPU scaling, and serving-ready tooling for production deployments.

This skill provides advanced guidelines for designing robust and effective tools specifically for autonomous AI agents. It shifts the focus from traditional developer APIs to contracts optimized for Language Models (LLMs). Key topics covered include treating tool descriptions as prompt engineering, implementing namespacing, applying the Consolidation Principle, and utilizing Architectural Reduction patterns (like file system access) to maximize agent reliability and workflow complexity.

Automates guidance on train/test split workflows inside ML training, surfacing best-practice patterns, code snippets, and validation tips for PyTorch, TensorFlow, and sklearn projects.

Provides production-ready Megatron-Core workflows for training 2B–462B parameter LLMs with tensor/pipeline/expert/context parallelism, maximizing GPU efficiency on H100 clusters.

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.

This skill provides a comprehensive workflow for integrating TwinMind meeting transcription services into continuous integration/continuous deployment (CI/CD) pipelines. It automates the processes of meeting documentation, real-time transcription (using Ear-3), and structured action item extraction using advanced AI models like GPT-4 and Claude. Use this when setting up automated meeting intelligence operations or managing enterprise-level transcription workflows via API.

This skill outlines security best practices for the TwinMind AI meeting assistant. It covers crucial aspects like on-device audio processing, encrypted cloud backups, and managing microphone permissions. Use it when setting up new TwinMind instances or managing AI operations to ensure data privacy, compliance, and robust security for sensitive meeting data.

This comprehensive workflow guides users through the end-to-end process of utilizing cloud-based GPU computing resources on Vast.ai. It covers searching for ideal GPU offers based on specs and price, provisioning an instance, running complex training or inference jobs via SSH, collecting model artifacts, and crucially, destroying the instance to prevent unexpected billing. Ideal for deep learning training and research.

This comprehensive script suite provides methods to optimize GPU cloud computing expenses on Vast.ai. It guides users through selecting the most cost-efficient GPU for their workload, leveraging interruptible (spot) instances, and implementing critical lifecycle management safeguards. Features include detecting idle compute resources, auto-destroy timers, and calculating total cost burn rates, ensuring AI model training and inference jobs stay within budget.

Automates Verifiedemail workflows through Composio's Rube MCP, guiding tool discovery, connection verification, and multi-tool execution while enforcing schema compliance and session reuse for reliable automation.

Provides automated guidance for designing, validating, and deploying Vertex AI pipelines within Google Cloud, covering compute, storage, BigQuery, and GCP-specific services for production readiness.