Settings and links
Odd settings: check:settings
Most settings are the same on nearly every course: sequential navigation, registration rules, progress tracking. A course that differs is usually a dashboard slip. check:settings finds those, read-only.
For each course, published course, path and published path setting, it finds the value most records use. If one value covers at least 90% of them, every record with another value is listed. Settings that are genuinely split, such as hidden, are reported as varying rather than flagged. Nothing is hardcoded, so the norm is whatever your catalog does.
syncjar check:settings # report
syncjar check:settings --domain courses.example.com # one domain; repeatable
syncjar check:settings --threshold 0.8 # a looser norm
syncjar check:settings --fail-on-outliers # exit 1 if anything is flagged, for CI
Domains come from --domain, then SKILLJAR_DOMAINS, then every domain the key can see. It uses the v1 API key.
Broken links: check:links
npm run check:links
Reads every lesson in the local preview, collects every link, and sends each unique URL a HEAD request with a five-second timeout. Anything that doesn’t answer 200 is written to public/data/link-report.json with the lessons that use it. Fragment (#…) and mailto: links are skipped.
It reads the preview’s copy of the content, so run npm run generate:courses first, and run it from a clone.
Some sites answer HEAD with an error while working fine in a browser, so treat the report as a list to look at rather than a list to fix.