Loading...
Loading...
Found 418 Skills
Complete FFmpeg + OpenCV + Python integration guide for video processing pipelines. PROACTIVELY activate for: (1) FFmpeg to OpenCV frame handoff, (2) cv2.VideoCapture vs ffmpeg subprocess, (3) BGR/RGB color format conversion gotchas, (4) Frame dimension order img[y,x] vs img[x,y], (5) ffmpegcv GPU-accelerated video I/O, (6) VidGear multi-threaded streaming, (7) Decord batch video loading for ML, (8) PyAV frame-level processing, (9) Audio stream preservation with video filters, (10) Memory-efficient frame generators, (11) OpenCV + FFmpeg + Modal parallel processing, (12) Pipe frames between FFmpeg and OpenCV. Provides: Color format conversion patterns, coordinate system gotchas, library selection guide, memory management, subprocess pipe patterns, GPU-accelerated alternatives to cv2.VideoCapture. Ensures: Correct integration between FFmpeg and OpenCV without color/coordinate bugs. See also: ffmpeg-python-integration-reference for type-safe parameter mappings.
Deploys and optimizes AI/ML inference workloads on GKE, using GPUs, TPUs, and model servers. Use when deploying GKE inference servers, configuring GKE GPU resources for inference, or deploying LLMs on GKE. Don't use for generic batch jobs or HPC task queues (use gke-batch-hpc instead).
Read-only Jetson health snapshot for identity, memory, GPU, thermal, power, storage, services, and top processes.
Manage CompShare GPU cloud resources through the compshare CLI. Use when Codex needs to install or configure compshare-cli, search GPU specifications and inventory, inspect pricing, create or manage instances, connect over SSH, transfer files, run durable remote jobs, manage images, disks, US3, teams or billing, ask product questions, or diagnose CLI problems.
The normative Design Guides for GPUI Kit desktop applications. Read in full before designing or changing any screen, layout, spacing, visual hierarchy, color, density, component choice, interaction state, overlay, motion, data-heavy view, or interface copy in a GPUI Kit (gpui-kit / gpui-component) application, and before reviewing UI work for design quality. Also use when asked what the design rules are, or whether a UI decision follows them.
Layout and styling in GPUI. Use when styling components, layout systems, or CSS-like properties.
Global state management in GPUI. Use when implementing global state, app-wide configuration, or shared resources.
Write, explain, and debug Metal GPU code including shaders (vertex, fragment, compute), render pipelines, compute pipelines, buffer/texture management, and Metal 4 APIs. Use this skill whenever the user mentions Metal, GPU programming, shaders, MSL (Metal Shading Language), render passes, compute kernels, MTLDevice, MTLCommandBuffer, MTLRenderPipelineState, or any Apple GPU/graphics programming topic. Also trigger when the user wants to do parallel computation on Apple devices, write GPU-accelerated code, or work with Metal Performance Shaders, MetalFX, MetalKit, or Compositor Services. Covers iOS, macOS, iPadOS, tvOS, and visionOS.
Iteratively optimize cuTile kernel performance through systematic profiling, bottleneck analysis, IR comparison, and targeted tuning. Covers tile sizes, occupancy, autotune configs, TMA, latency hints, persistent scheduling, num_ctas, flush_to_zero, and IR-level debugging. Use when asked to "optimize cutile kernel", "improve kernel perf", "tune cutile performance", "make kernel faster", or iteratively benchmark and refine a cuTile GPU kernel in the TileGym project.
DGX Cloud Lepton managed GPU compute platform with run/status/cancel interface. Use when submitting TAO jobs to DGX Cloud, dispatching training/eval/inference to Lepton GPU resources, or managing Lepton workspace deployments. Trigger phrases include "run on Lepton", "submit to DGX Cloud", "Lepton job", "managed GPU on DGX Cloud".
Auto-scale LLM inference clusters on Kubernetes using KEDA, custom GPU metrics, and horizontal pod autoscaling. Handle traffic spikes, implement queue-based scaling, and optimize cost with spot instances for AI workloads.
Optimizes transformer attention with Flash Attention for 2-4x speedup and 10-20x memory reduction. Use when training/running transformers with long sequences (>512 tokens), encountering GPU memory issues with attention, or need faster inference. Supports PyTorch native SDPA, flash-attn library, H100 FP8, and sliding window attention.