Skill UI

Activation-aware weight quantization that compresses 7B-70B models to 4-bit with minimal accuracy loss, delivering ~3x inference speedups and memory reductions for constrained GPU deployments while integrating with vLLM and Marlin kernels.

Provides a minimal Deepgram transcription sample covering client initialization, URL/local audio transcription, and error handling tips so you can test setups or learn API basics quickly.

Minimal Gamma integration sample showing how to initialize the SDK, create a simple three-slide presentation, and handle common setup errors; ideal for smoke testing new Gamma projects or learning basics.

Collects environment details, trace logs, and minimal reproduction scripts for LangChain issues, packaging them into a debug bundle that accelerates bug reports, diagnostics, and support conversations.

Provides a minimal LangChain example that walks through defining a prompt, chaining it to an LLM, and invoking it from Python or TypeScript to confirm the integration setup and learn basic chain/prompt patterns.

Provides a minimal Langfuse trace example that initializes the client, records spans and generations, flushes data, and logs a trace URL—ideal when starting integrations, validating setup, or learning trace patterns.

Provides a minimal Lindy AI agent example for starting a new integration, testing your setup, or learning core API patterns, guiding you through file creation, client initialization, and first agent execution.

nanoGPT is a simplified, minimalist implementation designed specifically for educational purposes. It reproduces the core architecture of GPT-2 (124M) using clean, hackable code, allowing users to understand the entire transformer pipeline from scratch. The framework supports full workflows, including data preparation on various datasets (e.g., Shakespeare, OpenWebText), model training, and text generation, making it ideal for students and researchers learning NLP and deep learning principles.

Provides a minimal Perplexity SDK example for bootstrapping a new integration, verifying credentials, and confirming that the API client can connect before building more advanced flows.

Provides a minimal Speak integration to showcase AI-powered language lessons, covering client setup, lesson creation, tutor prompts, feedback, and common setup errors during onboarding.

Generates minimal macOS Seatbelt sandbox profiles with allowlist-based resource rules so apps can be isolated by permissions, useful for restricting file, network, and IPC access when defending against high-risk execution or supply chain threats.

Aws Agentic Ai refers to an AI-driven agentic capability within AWS, aiming to automate complex workflows by orchestrating services and responding to evolving goals with minimal human intervention.