A node that provides a polygon shape to a [CollisionObject2D] parent. A node that provides a thickened polygon shape (a prism) to a [CollisionObject2D] parent and allows to edit it. The polygon can be concave or convex. This can give a detection shape to an [Area2D] or turn [PhysicsBody2D] into a solid object. [b]Warning:[/b] A non-uniformly scaled [CollisionShape2D] will likely not behave as expected. Make sure to keep its scale the same on all axes and adjust its shape resource instead. Collision build mode. Use one of the [enum BuildMode] constants. If [code]true[/code], no collisions will be detected. If [code]true[/code], only edges that face up, relative to [CollisionPolygon2D]'s rotation, will collide with other objects. [b]Note:[/b] This property has no effect if this [CollisionPolygon2D] is a child of an [Area2D] node. The margin used for one-way collision (in pixels). Higher values will make the shape thicker, and work better for colliders that enter the polygon at a high velocity. The polygon's list of vertices. Each point will be connected to the next, and the final point will be connected to the first. [b]Note:[/b] The returned vertices are in the local coordinate space of the given [CollisionPolygon2D]. [b]Warning:[/b] The returned value is a clone of the [PackedVector2Array], not a reference. Collisions will include the polygon and its contained area. In this mode the node has the same effect as several [ConvexPolygonShape2D] nodes, one for each convex shape in the convex decomposition of the polygon (but without the overhead of multiple nodes). Collisions will only include the polygon edges. In this mode the node has the same effect as a single [ConcavePolygonShape2D] made of segments, with the restriction that each segment (after the first one) starts where the previous one ends, and the last one ends where the first one starts (forming a closed but hollow polygon).