research-gapslisted

# /research-gaps — fill what nothing covers Takes the `Status: gap` subtopics of `catalog/topics.md` as its work list, researches each one, and writes what it finds into `knowledge/` as a source of its own — so the catalog entry stops being a gap and `/cards` picks it up on the next run. The name is `research-gaps`, not `research`: this plugin ships into environments that may already have a skill called `research`. **This is the only step that reaches the network**, and the only one that puts material into the project the user did not choose. Both facts shape everything below. ## Steps 1. No `catalog/topics.md` → point at `/catalog`, done. 2. Collect the subtopics marked `Status: gap`. **If there are none, say so and exit without touching anything.** 3. Without arguments, work the whole list; with arguments, only the gaps that match. 4. Per gap: search the web, read what you find, and write one synthesised document. Follow the `depth` in `goal.md` — `awareness` wants a page, `expert` wants the trade-offs — and aim at the 3–8 cards a subtopic normally yields, not an exhaustive survey. 5. Register the source, write the document, update the catalog entry. 6. Report what closed and what did not. ## What gets written **One source per run** in `sources.yaml`, or one per gap if the subjects are unrelated: ```yaml - id: <gap-slug>-research type: research gap: 'The exact subtopic name from catalog/topics.md' note: 'Synthesised for a catalog gap' ``