PCG

Split by whether they change your project. Read-only tools are always safe to try.

Reading

• Describe PCG Graph — Overview of a PCG graph: how many nodes, how they are wired, and which are unconnected. Start here before changing a graph.
• List PCG Components — PCG components placed in the current level, which graph each runs, and whether it is generating. Use this to find which actor a graph is actually driving.
• List PCG Nodes — Nodes in a PCG graph with their pins and, crucially, the settings_object path. Node properties live on that object - read and write them with hf_get_property / hf_set_property, using a dotted path for members nested in a struct.

Writing

• Connect PCG Nodes — Wire one PCG node's output pin to another's input pin, and VERIFY the edge exists afterwards. PCG drops an edge silently when a pin label does not match, so an unverified connect is worthless. Use 'input' or 'output' to reach the graph's own pins.