NikiforovAll sdlc-ontology
The model, and where each part of it is written down. This skill holds no procedure: mapping-session interviews a team into a document, auto-mine-repo reads one out of a repository, and both read the vocabulary from here so it is stated once.
If someone asks you to draw an arrow, the answer is to name an artifact. Everything on the page is either authored or derived, and the derived things are the ones that look most like drawings.
Where to read
| Reach for | When |
|---|---|
references/ontology-cheatsheet.md |
writing YAML — every field, every layer, and what check and status catch, on one page |
references/worked-example.md |
a level or a fill-versus-recommendation call will not settle; read it against ai-sdlc example |
docs/ontology/ |
maintaining the model — six documents, an overview plus five layers in dependency order. In the ai-sdlc repository only; the installed skill does not carry it |
You are done consulting this when every activity you have written is in exactly one of the three states — filled, open, or unclaimed — and you can say which, and why.