Geometry
Skill and tool bundles for gap (graph as policy) — Anthropic Agent Skills format, discovered by path
npx -y skills add graph-robots/open-robot-skills --skill geometryAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
What its author says it does
Copied from the file, not written here
Pure-math 3D geometry toolbox — back-project masks and depth to point clouds, DBSCAN-filter noise, fit oriented bounding boxes, derive top-down/front grasp poses, and reconstruct collision worlds from RGB-D frames. Use when a workflow needs perception geometry or planner inputs computed on CPU with no model weights.
The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.
SKILL.md
4.9 KB, as published. Nobody here has run it
geometry
Pure-math perception/planning geometry as in-process typed tools, from
mask back-projection through OBB fitting to grasp-candidate generation,
plus the two scalar helpers (geometry.iou, geometry.pose_distance).
Fully CPU — no model weights, no GPU.
When to use
- Turning a segmentation mask + depth + camera calibration into world-frame
points (
mask_to_world_points) and an object OBB (filter_and_compute_obb). - Deriving grasp poses from an OBB:
top_down_grasp_candidatesfor tabletop pick (feed the full list tocurobo.plan_to_grasp_posesas a goalset),front_grasp_from_obbfor horizontal interactions (drawer/door handles). - Building the collision world for the planner:
build_world_configwith the target's mask inobject_masksso the planner canignore_obstacle_namesit.
Install
uv sync --extra geometry # open3d + scikit-learn (cv2/scipy come with gap core)
# (pip: pip install -e ".[geometry]")
The module imports lazily — the bundle loads (and the light tools work) without the extra; only OBB fitting, DBSCAN filtering and world reconstruction need open3d/sklearn/cv2.
Gotchas (carried over from the service)
- OBB
extentis HALF-extents (gap.types convention, same as the proto).compute_obbis upright-only: rotation is around world Z (no 3D tilt), and extents use the 2nd/98th percentile of points, not strict min/max. - Single-camera clouds are 2.5D: only camera-facing surfaces are observed, so
OBB centers carry a few cm of depth bias on opaque objects. (The service's
rehearsal-sandbox ground-truth snap that compensated for this in-container
was deliberately NOT ported — it depended on a
/appsandbox file.) top_down_grasp_candidatesdefaultz_offset=-0.04: fingertip 4 cm below the OBB top. Withz_offset=0.0the fingers close above the object (silent empty grip). Grasp Z is clamped to -0.05 m (table-clearance floor; LIBERO table top is at world z=0).mask_to_world_pointskeeps only depths in [0.015, 20.0] m (HyRL bounds); invalid/zero-depth pixels are dropped.filter_noisereturns the ORIGINAL cloud unchanged when DBSCAN labels everything noise (defensive fallback, mirrors HyRL).build_world_config: table removal only runs whentable_z_threshold != 0(typical -0.01); robot-point exclusion is Franka-only (simplified DH FK) and skips non-7-DOF joint states; prefer explicitobject_masksover thetarget_obbprojection fallback — masks are pixel-accurate, the OBB projection is a corner-AABB approximation inflated by 2 cm.top_down_grasp_from_obbyaw is NOT derived from the OBB — fingers may close across the wide axis; use the candidate fan when orientation matters.