Physical AI expertise for coding agents

Entry-point skill for shaping new robotics applications: lightweight brainstorming, project ideation, requirement/risk discovery, stack selection, comparing or choosing simulators/models/visualization, a first user-visible slice, and a concise architecture decision record.

Modern Gazebo (gz, Harmonic/Ionic line) simulation: SDF worlds and models, sensors (lidar, camera, IMU, contact), the ros_gz bridge, spawning robots, and headless/server operation.

NVIDIA Isaac Lab: reinforcement-learning and imitation-learning workflows on top of Isaac Sim: prebuilt environments and tasks, training runs, and exporting policies.

NVIDIA Isaac Sim: installation and container setup, GPU/driver requirements, USD scenes, robots and sensors, the ROS 2 bridge, and headless/livestream operation for remote servers.

MuJoCo for lightweight, contact-rich robot manipulation simulation on macOS/Linux, especially single-arm grasping without ROS: headless offscreen rendering, MJCF models, mujoco_menagerie assets, damped-least-squares inverse kinematics, and empirical grasp calibration.

Choose and set up robotics simulators, and simulate sensors correctly: Gazebo vs Isaac Sim selection, sensor fidelity (rates, noise models, frames matching the real robot), determinism, and sim-to-real considerations.

Data sourcing strategy for robotics and physical-AI: choose between offline datasets (HuggingFace hub, Open X-Embodiment and similar), simulation-generated data, and teleop/real-robot collection; plan storage formats, episode structure, and dataset versioning.

Hugging Face Hub operations for robotics projects: inspect, download, create, upload, authenticate safely, explore Dataset Viewer data, run and diagnose Jobs, and inspect Spaces.

HuggingFace LeRobot for physical-AI manipulation: the LeRobotDataset format, loading and recording episodes, training policies (ACT, diffusion, pi0) and VLAs (SmolVLA), inspecting checkpoint processors, evaluating in simulation, LeRobot training-job lifecycle, and teleoperation.

Foxglove and Lichtblick for robotics visualization: foxglove_bridge setup for live ROS 2 robots, choosing the open-source Lichtblick client versus Foxglove, layouts, MCAP playback/recording, and remote web visualization.

Rerun for data-centric robotics and ML visualization: logging APIs (Python), timelines, entity paths, and viewing policy rollouts, episode data, and sensor streams.

RViz2 visualization for ROS 2: displays, TF frame debugging, markers, saved config files, and the common 'nothing shows up' fixes (fixed frame, QoS, sim time).

Choose and apply robotics visualization: selection guidance for rviz2 vs Foxglove vs Rerun, plus best practices: what to visualize at each dev stage, live vs recorded, local vs remote.

Virtual-environment-first setup for robotics projects: decide uv/venv vs Docker, make local and remote-server runs reproduce identically, handle GPU passthrough and headless/display forwarding.

Glue robotics modules into one running system: choose module boundaries, pick inter-module communication (ROS 2 topics/services/actions, zenoh, gRPC, REST, shared memory), and write solid Dockerfiles and docker-compose for robotics workloads.

Nav2 mobile-robot navigation for ROS 2: bringup, behavior trees, costmaps, planner/controller servers, localization (AMCL, slam_toolbox), waypoint following, and tuning.

Core ROS 2 usage: workspaces, colcon builds, packages (ament_python/ament_cmake), nodes, topics/services/actions, QoS, launch files, parameters, TF2, rosdep, and gluing third-party packages together.

Deploy headless robotics / sim / demo containers to Google Cloud Run: the build → Artifact Registry → Cloud Run path plus the gotchas that bite sim workloads (no UDP multicast for gz-transport/DDS, CPU allocated only while a request is open, session affinity for per-visitor instances, request timeout / concurrency for long-lived WebSockets, VPC subnet sizing).

Add safe hosted-session runtime to a working robium app: per-visitor instances, a capability-protected gateway, start/status/stop lifecycle, boot logs, fleet budgets, Cloud Run or RunPod drivers, and a Gradio, Lichtblick, or Foxglove viewer handoff.

Publish an implemented, smoke-tested Robium application as one coherent public project: write or revise its product-lab article, prepare real-run media and explanatory assets, apply the Robium editorial cobalt theme, and compose catalog, overview, live, guide, and social surfaces from repository-owned facts.

Robium's explicit learning engine: consolidate capture-only queue flags and transcript evidence into observations; absorb ready observations through deterministic anchor-targeted skill deltas; refine, A/B contested edits, deep-verify examples, prune landed transcript evidence, and report loop health.

Registry-driven mining of external example repos: the learning engine's second experience source. Surveys, deep-reads, and comparatively analyzes approved repos (vendor demos, framework samples, community robot apps) into evidence-cited observations (origin: external) that harden robium skills or propose new ones; maintains crawl records for drift re-checks.

Operate RunPod Pods safely: inspect balance, compare alternative GPU resources and US regions, launch immutable images, attach network volumes, expose services, diagnose logs/startup, validate workloads, and clean up paid compute.

Author new robium skills and enforce the catalog quality bar. Owns the authoring workflow from templates/skill, the quality bar (references/quality-bar.md: template compliance, trigger-surface descriptions, <500-line bodies, stated delegation posture, upstream links, no invented syntax), and scripts/validate_skills.py: run it after ANY skills/ change.

Canonical test assets and fixture sourcing for robotics testing: which worlds, robot models, sample datasets, and recordings to test against for a given robot type; the standard test-assets folder layout with a provenance catalog and per-asset manifests; pointer vs vendored sourcing modes; fixture and golden policy (tolerance bands, seeded generation).

Test-driven robotics development: smoke tests for launch files, sim-based regression tests, node-level unit tests, policy eval as a test, and CI patterns for robotics repos.