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
teleoperation-and-collectionGuides interactive RoboCasa keyboard or SpaceMouse teleoperation, demonstration collection, safe interruption recovery, and raw HDF5 handoff boundaries.
Independent GitHub snapshot · Automated classification · Original documentation governs
Use this sub-skill when a task asks to control a RoboCasa robot interactively or record successful human demonstrations. Both workflows open a viewer and consume live input; collection also writes episode files and mutates HDF5 output. Never present them as unattended or automatically verified.
pynput/hidapi failures, SpaceMouse ID mismatches, missing assets, partial collection, unwritable output, or conversion failures.scripts/check_teleop_prereqs.py before an interactive session. Its default mode only inspects environment variables and package metadata; it never opens a viewer, writes a dataset, or enumerates HID devices.simulation-environments sub-skill.datasets-demonstrations sub-skill.tasks-scenes-assets sub-skill.setup_macros.py is an interactive, potentially overwriting helper. Read the warning in teleoperation.md; do not run it automatically.--directory with enough space and decide whether raw episode directories must be retained.--help or the bundled diagnostic non-interactively. Mark live teleoperation and playback as deferred interactive evidence.