Use when routing an ambiguous Computer Vision task to a bounded local, webapp, model, dataset, demo, and verification skill stack.
Use when starting a Computer Vision web application with Next.js, FastAPI, browser MediaPipe, typed APIs, bounded uploads, and private media handling.
Use when extracting, labeling, splitting, converting, validating, or versioning Computer Vision datasets with provenance, privacy, and leakage controls.
Use when building browser or Python hand, gesture, face, or pose interactions with smoothing, semantic events, and privacy-preserving media handling.
Use when building or reviewing OpenCV image, video, or webcam frame loops with deterministic fallbacks, performance measurement, and reliable cleanup.
Use when packaging an existing Computer Vision pipeline as a reproducible local CLI or lightweight demo with explicit artifacts and verification.
Use when verifying Computer Vision imports, fixtures, models, schemas, cameras, artifacts, cleanup, latency, and stated limitations with reproducible evidence.
Use when designing or verifying Ultralytics YOLO detection, segmentation, tracking, or pose inference with licensed models and stable JSON outputs.