| name | mira-light-orchestrator |
| description | Safely orchestrate Mira Light scenes, status checks, and low-level recovery through the local bridge. |
| metadata | {"openclaw":{"requires":{"config":["plugins.entries.mira-light-bridge.enabled"]}}} |
Mira Light Orchestrator
Use this skill when the task involves Mira's physical body, scene choice, lamp
state, or the boundary between vision events and motion.
Default workflow
- Read
mira_light_runtime_status and mira_light_status if physical state matters.
- Prefer
mira_light_list_scenes or known scene names to reason about behavior.
- Use
mira_light_run_scene for embodied responses.
- Use
mira_light_set_led for light-only adjustments.
- Use
mira_light_stop or mira_light_reset for recovery.
- Use
mira_light_control_joints only when the human explicitly wants
calibration, direct rehearsal, or recovery from a bad pose.
Scene-first policy
The body should usually express intent through named scenes rather than raw
servo commands. Scene-first keeps the behavior legible and preserves Mira's
personality.
Good defaults:
- greet or first notice:
wake_up
- gentle attention:
curious_observe
- warmth or contact:
touch_affection
- playful hesitation:
cute_probe
- ambient aliveness:
daydream
- positive payoff:
celebrate
- departure:
farewell
- rest:
sleep
Vision policy
When the task mentions seeing, target tracking, or camera input:
- treat vision outputs as event data
- look for
scene_hint and control_hint
- prefer routing through the runtime bridge instead of inventing servo angles
- remember that
track_target is still experimental
Reference files:
{baseDir}/../../TOOLS.md
__REPO_ROOT__/config/mira_light_vision_event.schema.json
__REPO_ROOT__/scripts/track_target_event_extractor.py
__REPO_ROOT__/scripts/vision_runtime_bridge.py
Recovery rules
- If the bridge or lamp seems unhealthy, stop and inspect before sending more motion.
- If a user request conflicts with safety or scene readiness, explain the safer
alternative and propose the closest scene.
- Preserve the feeling of intention: small safe actions are better than large unsafe ones.