Ros1 Development

If you're maintaining legacy ROS1 systems or haven't migrated to ROS2 yet, this gives you the architectural patterns that prevent common disasters. It walks through node decomposition (avoiding monolithic "do-everything" nodes), proper queue sizing to prevent memory leaks, tf tree management, and actionlib patterns with preemption handling. The launch file examples show conditional loading and proper namespacing. Most valuable for the specific footguns it warns against: queue_size=0 on high-frequency topics, timestamp comparison bugs, and transform lookup without exception handling. Covers the full catkin workspace ecosystem including nodelets and the ROS1-to-ROS2 migration path when you're finally ready to jump ship.

GitHub stars: 265
last updated: May 25, 2026

View on GitHub

Install

$ npx -y skills add arpitg1304/robotics-agent-skills --skill ros1-development --agent claude-code

Installs into .claude/skills of the current project.

Files

ROS1 Development Skill

When to Use This Skill

Building or maintaining ROS1 packages and nodes
Writing launch files, message types, or services
Debugging ROS1 communication (topics, services, actions)
Configuring catkin workspaces and build systems
Working with tf/tf2 transforms, URDF, or robot models
Using actionlib for long-running tasks
Optimizing nodelets for zero-copy transport
Planning ROS1 → ROS2 migration

Core Architecture Principles

1. Node Design

Single Responsibility Nodes: Each node should do ONE thing well. Resist the temptation to build monolithic "do-everything" nodes.

# BAD: Monolithic node
class RobotNode:
    def __init__(self):
        self.sub_camera = rospy.Subscriber('/camera/image', Image, self.camera_cb)
        self.sub_lidar = rospy.Subscriber('/lidar/points', PointCloud2, self.lidar_cb)
        self.pub_cmd = rospy.Publisher('/cmd_vel', Twist, queue_size=10)
        self.pub_map = rospy.Publisher('/map', OccupancyGrid, queue_size=1)
        # This node does perception, planning, AND control

SKILL.md on GitHub