Pywayne Visualization Rerun Utils

Static 3D visualization utilities wrapping Rerun SDK for adding point clouds, trajectories, cameras, planes, and chessboards. Use when visualizing 3D data in...

Security Scan

The skill's stated purpose (static 3D visualization helpers for the Rerun SDK) matches its instructions and requirements, but it omits an explicit install spec and relies on an unspecified 'gettool' auto-download — verify the dependency source before use.

Purpose & Capability

Name and description match the SKILL.md content: helpers for adding point clouds, cameras, planes, etc., to a Rerun viewer. The declared needs (Rerun SDK) are coherent with the stated purpose.

Instruction Scope

Instructions are narrowly scoped to visualization tasks (calling rr.init, static add_* methods, data formats). They do not request unrelated files, credentials, or system-wide configuration. However, SKILL.md says the Rerun SDK is 'auto-downloaded via gettool' — that is an operational detail that can cause external package installation and should be clarified (which package name, version, and source).

Install Mechanism

There is no declared install spec; the SKILL.md relies on a platform helper ('gettool') to auto-download the Rerun SDK. That is plausible for this purpose but leaves the install source/verification unspecified (no package name/version or trusted release host). This is a moderate procedural risk to confirm before installation.

Credentials

The skill declares no required environment variables, credentials, or config paths and the instructions do not reference secrets or unrelated environment data. The requested privileges are proportionate to a visualization utility.

Persistence & Privilege

always is false and the skill does not request persistent system-wide changes or cross-skill configuration. Autonomous invocation is allowed by default (normal for skills) but does not combine here with other concerning privileges.

Assessment

This skill appears to do what it says: helper functions for Rerun visualization. Before installing or allowing the agent to auto-install dependencies, verify the Rerun SDK source: ask for the exact package name and version and confirm it comes from an official/known distribution (PyPI/GitHub releases). If possible, run the install in a sandboxed environment first. Also confirm what your platform's 'gettool' does (which registry it pulls from and whether it can run arbitrary install scripts). If you need stronger assurance, request the skill author to include a concrete install spec (package name/version and checksum or GitHub release URL) so you can validate the dependency prior to installation.

pywayne.visualization.rerun_utils.RerunUtils provides static methods for adding 3D elements to a Rerun viewer.

Quick Start

import numpy as np
import rerun as rr
from pywayne.visualization.rerun_utils import RerunUtils

# Initialize Rerun (called once globally)
rr.init('my_app', spawn=True)

# Use static methods to add elements
RerunUtils.add_point_cloud(points, colors=[255, 0, 0])
RerunUtils.add_trajectory(trajectory)
RerunUtils.add_camera(pose, image='path/to/image.jpg')

Point Cloud

# Single color (default: red)
RerunUtils.add_point_cloud(points)

# Multi-color points
colors = np.random.randint(0, 255, (100, 3))
RerunUtils.add_point_cloud(points, colors=colors, label='colored')

# Data format: points (N, 3)

Trajectory

# Single-color trajectory (default: green)
RerunUtils.add_trajectory(trajectory)

# Multi-color trajectory
colors = np.array([[0, 255, 0], [255, 0, 0]], dtype=np.float32)
RerunUtils.add_trajectory(trajectory, colors=colors, label='path')

# Data format: traj_endpoints (N, 3)

Camera

# Camera only (no image)
RerunUtils.add_camera(pose, label='main_camera')

# Camera with image
RerunUtils.add_camera(pose, image='path/to/image.jpg', label='rgb_camera')

# Data format: camera_pose (4, 4) or (4, 7)

Plane

# Plane by center and normal
RerunUtils.add_plane_from_center_and_normal(center, normal, half_size=1.0)

# Plane from SE3 transformation matrix
RerunUtils.add_plane_from_Twp(Twp, half_size=1.0)

Chessboard

# Standard chessboard
RerunUtils.add_chessboard_from_Twp()

# Custom chessboard with colors
RerunUtils.add_chessboard_from_Twp(
    rows=9, cols=6, cell_size=0.025,
    Twp=pose_matrix,
    color1=np.array([255, 0, 0]),  # Red
    color2=np.array([0, 0, 255])   # Blue
    label='calib_board'
)

Internal Helper

# Get quaternion from v1 to v2 (used internally for plane rotation)
quat = RerunUtils._get_quaternion_from_v1_and_v2(v1, v2)

Important Notes

• Initialization: Call rr.init('name', spawn=True) once before using methods
• Static methods: All methods are static class methods, no instance needed
• Dependencies: Requires Rerun SDK (auto-downloaded via gettool)
• Data types: All position inputs must be float32
• Coordinates: Rerun uses ViewCoordinates.RDF (robot-centric coordinate system)
• SE3 poses: Support (4, 4) or (4, 7) matrix formats
• Color format: RGB as numpy arrays with shape (3,)