executing-plans
executing-plans
Use when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available
video-world-streamingPlan and troubleshoot Sana video-generation, long-video, world-model, and streaming V2V workflows.
Independent GitHub snapshot · Automated classification · Original documentation governs
Use this sub-skill when a future agent needs to plan, compare, or debug Sana video workflows that share the same video-generation stack but differ in prompt surface, camera control, long-context rollouts, or streaming editing.
bidirectional_short and long_streaming video-to-video editing.image-generation.training-data-configs.evaluation-conversion-deployment..npy camera / intrinsics arrays without running any model.sana-video when the user wants a direct SANA-Video clip plan.sana-video-refiner when the user wants base SANA-Video plus LTX-2 refinement.sana-wm when the user wants camera-controlled world-model generation from an image, prompt, and action string or pose trajectory.sana-wm-streaming when the user wants the chunk-pipelined world-model path, including fp8/fp4 planning and progressive MP4 output.sana-streaming-v2v when the user wants source-video editing with long_streaming or bidirectional_short.a/d as yaw and j/l as strafe in the updated action mapping.fp8 and fp4 as precision choices that need explicit hardware awareness.a/d and j/l meaning.fp4 without Blackwell-class hardware.--action, --camera, or --intrinsics.