Adom Wiki
Public Made by Adomby adom
How the Adom Wiki (wiki.adom.inc) works and how to drive it with the adom-wiki CLI: the two storage layers (git page repo vs package registry), hosting source by pushing files (not a tarball), making
discover find/API: return metadata.activate inline (Gap-A from #47)
Gap-A (carved out of #47): the discovery search endpoints do not return metadata.activate inline, forcing a second fetch.
GET /api/discover and /api/discover/find return discovery_triggers, discovery_pitch, tags, star_count, vouch_count (see cleanDiscoverEntry), but NOT the activate recipe (metadata.activate: kind, skill, bridgeManifestUrl, postInstall, requires, verbPrefixes). So an AI that discovers a repo via discover find must do a SECOND round-trip to .../files/page.json just to read the activate recipe before it can install/activate.
Ask: return activate inline in the discovery results so discovery -> activate is one hop.
Why it is non-trivial (design note): activate lives in the repo page.json, not the indexed pages row, which is why the discovery scorer/endpoints (built to avoid N+1) do not carry it. Returning it inline means either (a) a per-result page.json read (a git file read per row -- the N+1 the discovery path deliberately avoids), or (b) indexing metadata.activate into a pages column at publish/index time (the cleaner fix -- add an activate_json column via addColumnIfMissing, populate it in indexRepo, and select it in cleanDiscoverEntry). Option (b) is the recommended path.
The Discover TAB (the headline of #47) already reads activate live per page, so it is unaffected; this is purely the API optimization from the "Gap-A" section of #47.