Skill UI

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.

Automates Textrazor tasks through Composio’s Rube MCP pipeline, using RUBE_SEARCH_TOOLS, RUBE_MANAGE_CONNECTIONS, and RUBE_MULTI_EXECUTE_TOOL so each workflow picks up the current schema and active toolkit connection.

Automates Tinyurl operations through Composio’s Rube MCP by guiding you to discover tool schemas, manage Tinyurl connections, and execute actions via RUBE_MULTI_EXECUTE_TOOL with session and memory handling.

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.

torchforge is Meta's advanced PyTorch-native library designed for agentic Reinforcement Learning (RL). It fundamentally separates core RL algorithms from complex distributed infrastructure concerns. This allows researchers to rapidly prototype and experiment with novel RL methods (like GRPO) using clean, PyTorch-native abstractions, while automatically handling large-scale tasks like distributed training across thousands of GPUs using Monarch and TorchTitan.

Automates generating TorchServe deployment configurations with guided best practices for model serving, MLOps pipelines, inference, and production monitoring so ML deployment workflows stay consistent and reliable.

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.

Automates Triton inference config guidance for ML deployment, covering serving, MLOps pipelines, monitoring, and production-ready configurations aligned with best practices.

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 guide provides comprehensive steps for deploying TwinMind integrations into production environments. It covers configuring the Chrome extension, setting up mobile app access, and managing API keys. TwinMind utilizes advanced AI models (including Ear-3 for transcription and GPT-4/Claude/Gemini for summarization) to provide real-time meeting transcription, summarization, and action item extraction. Ideal for professional teams managing high-volume AI meeting operations.