ignition-tag-authoringlisted

# Ignition Tag Authoring Ignition tags are the canonical data model that binds external sources (PLCs, databases, expressions) to the rest of the Gateway. This skill guides authoring, reviewing, and refactoring tag JSON so that what Copilot produces actually imports into Ignition Designer without silent failures. ## Critical precondition — ground truth **Before generating any tag JSON**, confirm the schema shape from a real Ignition export. Priority order: 1. A sample in the consumer project's `ground-truth/tags/` directory that matches the target tag shape (UDT Definition, UDT Instance, standalone AtomicTag, Folder). 2. A sample the user provided in the current conversation. 3. The verified schema catalog in [knowledge/ignition/tag-json-schema.md](../../knowledge/ignition/tag-json-schema.md), which documents what has been observed in existing ground truth. If the required shape is not covered by any of the above: - State clearly: "I'm emitting this with some fields inferred from Ignition 8.1 documentation but not directly verified against a sample of this kind. I'll flag those fields explicitly." - In the output, separate fields into **Verified** (seen in ground truth) and **Inferred** (derived from docs or general Ignition knowledge). - Ask the user for a matching export so the schema can be tightened on the next pass. Do NOT silently emit invented field names. This is the primary failure mode this skill exists to prevent. ## Key concepts — mental model ### Tag ana