OpenClaw · Skills: AI & machine learning

Patterns for running the Mapbox MCP server inside AI apps and agent frameworks.

Converts documents, images, audio, and web pages into Markdown for LLM processing.

NVIDIA guide for starting distributed Megatron-LM training on SLURM.

Media layer for HyperFrames: resolves music, effects, images, voice, and grades into local files.

NVIDIA skill for exporting sanitized evidence from medical AI inference to MLflow.

AWS skill for data warehouse migration to Redshift, currently supporting Teradata sources.

Summarizes per-model usage and cost for Codex or Claude from local CodexBar data.

Shows an on-screen alert with a message card and pulsing borders to get attention.

NVIDIA guide for picking and setting up distributed training strategies for AutoModel.

NVIDIA skill for configuring how NeMo AutoModel jobs launch: interactive, Slurm, or SkyPilot.

NVIDIA guide for adding new model architectures to AutoModel.

NVIDIA skill for creating and editing AutoModel recipes for training and evaluation.

NVIDIA skill for integrating NeMo Fabric into apps and platforms through its typed Python SDK.

NVIDIA skill for running Megatron-LM and Bridge training on mock or real data.

NVIDIA skill for turning single-node scripts into multi-node Slurm jobs and debugging them.

NVIDIA guide for trading compute for GPU memory with activation recompute on Megatron Bridge.

NVIDIA guide for validating and applying CPU offloading on Megatron Bridge.

NVIDIA guide for validating and applying CUDA graphs on Megatron Bridge.

NVIDIA guide for overlapping MoE expert-parallel communication on Megatron Bridge.

NVIDIA operational guide for turning on hierarchical context parallelism for Megatron Bridge.