2026-10-08From repairs to compilation: the site now has a single source of truth (content/*.json — 51 tools, 93 canonical topics with 40 alias cards, 16 expert profiles), a build gate that must pass before anything ships (data integrity, sitemap coverage, HTML tag balance, the quality checks and the claims ledger), a shared-partials drift sentinel, a self-generating review queue for the twelve-month re-review promise, and a live deploy self-check. The deploy check’s first run confirmed, with machine evidence, that the public site still serves an older build — a gap previously known only from manual probing. The allowlist in the claims ledger also caught and led to removal of a legacy link to an affiliated product on the practitioners page, enforcing the site’s own disclosure rules.
2026-10-08Attribution and performance pass: embedded assessments now carry first-party campaign parameters and a standard referrer policy (previously the embeds suppressed all origin information, making the referral pipeline unmeasurable while the site was publicly live), the Publish form records its source, and all content images below the fold load lazily. This changes no privacy posture: the site still runs no third-party analytics or trackers — attribution here is first-party only, between the reader’s browser and the delivery platform we operate.
2026-10-08Cross-page data audit: hub directory cards and entry pages disagreed on question counts and durations in 17 places (e.g. a mindset tool listed as 6 items in the hub and 12 on its page). Hub data now mirrors each entry page, and the quality checker refuses any future drift. One conflict remains visible by design — the Big Five Expert edition is described as 25 items on its page and 60 in the directory; until the live form is confirmed with the delivery team we neither hide it nor guess it. Storage copy on assessment embeds now promises only what AssessWiki itself guarantees (no answers stored by us); the delivery partner’s policy page is temporarily unreachable and we removed our pointer to it rather than cite a dead reference.
2026-10-08Content substance pass: several assessment entries shared near-identical “what it cannot tell you” and “what you take away” text (up to 90% word overlap between sibling tools). Eight entries — the three growth-mindset editions, both strengths reflection editions, the personal conflict edition, the general decision-style edition, and the student Big Five — were rewritten so each page states limitations and takeaways that are true of that instrument alone. The quality checker now fails any intra-cluster pair above 75% overlap, so the standard holds itself. The Q&A index reconciles its dual counts (133 answer cards covering 93 canonical topics), and embedded assessments gained context-aware heights.
2026-10-08Trust moved from wording to systems: the site's public promises are now machine-checked by an executable ledger (free library, review cadence, no trackers, named editor, counts), the quarterly checks cover all content directories and the assessment delivery pipeline, and error reports became one-click emails. Q&A cards gained a link to the researcher profile behind each framework, so every topic traces to a named source.
2026-10-08One library, one price: every assessment and every report on AssessWiki is free. Remaining paid-report wording was removed from the affected entry, Q&A, and zone pages — earlier copy had described a handful of deeper reports as paid, which no longer reflects how the library works. Entry review lines now read "Last updated" (what the date actually reflects), and the Policy commits to editorial re-review of every entry at least every twelve months, recorded here.
2026-10-08Theory Portals cold start, handled honestly: the empty directory state now invites authors to claim their portal instead of reading as a gap, and Publish explains the two listing routes (library submission, portal claim with authorship verification). AssessWiki does not create portal content for authors and never has — portals exist only when their real authors claim them.
2026-10-07Norms & interpretation series published: average EQ score, burnout scores, Big Five facet scores — alongside the existing personality norms page. All explain published reference data with its limits; AssessWiki publishes no norms of its own.
2026-10-07Citable references added: assessment entries, Q&A topics, expert profiles, comparisons, norms and glossary pages now carry a "How to cite this page" block (APA, MLA, HTML) pointing at the page's canonical URL. A site-wide citation and reuse guide was published.
2026-10-07Every assessment entry now shows its review line — "Last reviewed · Reviewed by KeoFung (AssessWiki Editorial)" — matching what the Q&A, comparison and norms pages already carried.
2026-10-07Accessibility and print pass: skip-to-content links and visible focus styles on every page; a shared print stylesheet so content pages print cleanly. Page sharing (link, X, LinkedIn, WhatsApp) added at page level only — assessment answers never leave the session and are never part of any share.
