fix(manifest): explicit embodiment_tags ["any"] (ADR-0071)

Empty embodiment_tags is no longer valid in the OpenRAL loader; embodiment-agnostic skills declare the explicit "any" wildcard.

SKILL.md

@@ -9,6 +9,7 @@ metadata:
   manifest: ./rskill.yaml
   role: s1
   kind: detector
+  embodiment_tags: [any]
   actions: [detect]
   objects: [person, cup, bottle, bowl, chair, table]
   scenes: [tabletop, kitchen, indoor]
@@ -41,6 +42,7 @@ An OpenRAL **object detector** (`role: s1`, `kind: detector`). RT-DETR-L (Real-T
 - **Verbs:** detect
 - **Objects:** person · cup · bottle · bowl · chair · table
 - **Scenes:** tabletop · kitchen · indoor
+- **Embodiments:** any
 
 ## Why this is discovery-only
 

rskill.yaml

@@ -19,26 +19,11 @@ role: "s1"
 kind: "detector"  # ADR-0037: perception producer; emits ObjectsMetadata, no Action
 
 # ── Compatibility contract ──────────────────────────────────────────────────
-# A detector is camera-based, not manipulator-specific. All embodiments are
-# declared because the only requirement is an RGB camera stream.
-embodiment_tags:
-  - "aloha"
-  - "franka_panda"
-  - "g1"
-  - "google_robot"
-  - "gr1"
-  - "h1"
-  - "mobile_base"
-  - "openarm"
-  - "panda_mobile"
-  - "pusht"
-  - "rizon4"
-  - "sawyer"
-  - "so100_follower"
-  - "so101_follower"
-  - "ur10e"
-  - "ur5e"
-  - "widowx"
+# embodiment-agnostic: a perception rSkill (detector/vlm) needs only an RGB
+# camera, not a specific robot. The rSkill<->robot gate exempts these kinds
+# (loader._EMBODIMENT_AGNOSTIC_KINDS, ADR-0037), so the tag list is empty —
+# it matches any robot/scene.
+embodiment_tags: ["any"]  # explicit embodiment-agnostic wildcard (ADR-0071)
 
 # The detector reads any RGB camera stream. vla_feature_key is omitted so
 # the loader accepts any camera key, not just "camera1".