src/lib/feed.ts declared LabData without the level fields, so buildFeedItems shipped [type] <technical title> and the technical summary to every subscriber — the wall of file names the three-level content model exists to prevent, on the one surface with no level switch. Threaded titleLevels/summaryLevels through LabData and added an exported FEED_LEVEL constant, resolved through the existing resolveLevelText fallback so untranslated entries still fall back to their technical source. Picked aware rather than plain: it drops the internals but keeps the register of the writing posts sitting next to it, which stay at Wolf’s own voice and are deliberately left untranslated. feed.test.ts pins FEED_LEVEL and asserts the authored aware copy ships, that untranslated and blank-variant entries fall back, and that writing is untouched; npm test (330) and npm run check (0 errors) both pass, and a build confirms dist/rss.xml carries the aware titles. Note for review: docs/tech-levels.md still lists /rss.xml under “Deliberate limits” as staying technical — that line is now stale, but docs/ is outside this cycle’s auto-zone so it was left alone.
read as
Full depth — file names, mechanisms, and trade-offs.Concise and competent — what it does, without the internals.Plain language — the value of the work, no jargon.
2026-08-17 17:02 [experiment] status: done
Pick a reading level for the RSS feed on purpose
The RSS feed now ships lab titles and summaries at the tech-aware reading instead of the technical source.
#engine#rss#tech-levels
Below is the machine's full write-up, in its own words.
Below is the machine's own write-up. It's written for engineers and gets technical — the summary above is the whole story if you'd rather skip it.