elk-listing-evaluation

IN BUILD

"the verdict was to either widen the offering to any kind of listing kit on etsy (no embroidery hoops) or shift the prism a bit to offer an evaluation of existing listings along with the fixed assets right in the same ui. So the verdict was to pivot, basically." (founder, 2026-08-23) … "its the earlier discovery for shifting the prism - lets proceed with 2" (founder, 2026-08-30)

STAGES

  1. proposal08-30 ↻ 08-31
  2. requirements08-30 ↻ 08-31
  3. design08-30 ↻ 08-31
  4. tasks08-30
  5. build—
  6. archived—

CAPABILITIES — 2 IN ONE CHANGE

  • listing-evaluation4 requirements
  • paid-asset-pack4 requirements

OUTCOMES — AND HOW EACH IS MEASURED

  • ○1.1A public listing URL returns a rendered evaluation (current-state card: identity, required pass/fail, recommended %, field checklist)not stated
  • ○1.2Every recommendation carries its evidence (report cards: photo slots, character blocks, alt-text status, video absence)not stated
  • ○1.3Recommendations cite Etsy's current documentation (quote + source + "checked <date>" from the citation registry)not stated
  • ○1.4The free evaluation includes one sample kit suggestion (title card's "What the kit would suggest")not stated
  • ○1.5Recommendations speak in opportunities (framing what's available, facts stated plainly beneath)not stated
  • ○1.6Full image slots with weak photos still get an image recommendation ("Improve the listing images you already have" with resolution chips)not stated
  • ○1.7A fully populated listing gets opportunities, not invented gaps (100%, "Nothing to fill — now make it work harder", test/refresh cards)not stated
  • ○1.8A shop link suggests its featured listing (confirm card, nothing scored until confirmed)not stated
  • ○1.9An unusable URL fails honestly (plain-language reason, nothing scored or stored)not stated
  • ○1.10The same listing data scores identically everywhere (ELK evaluation and labs scorecard share one rubric module)not stated
  • ○1.11Evaluation events fire without PII (evaluation-start / evaluation-complete / pack-offer-click, listing id only)not stated
  • ○1.12The report resolves into the kit offer (per-card "What's in the kit" boxes + the full-kit card, video as coming soon)not stated
  • ○1.13The kit delivers suggestions, not writes (ten images + template + title + tags + alt text as paste-ready output; no Etsy writes)not stated
  • ○1.14No embroidery framing survives anywhere (landing, evaluation, upload, result, emails)not stated

WHAT HAPPENED

  1. ↺ BACK TO DESIGN

    proposal and specs and design and tasks landed

    2026-08-30 · 4 pull requests
    • #414docs(openspec): propose elk-listing-evaluation
    • #415feat(elk): evaluation foundations — rubric, citations, endpoint
    • #416feat(elk): evaluation surface — the pivot's new front door
    • #417fix(elk): sell only what the kit ships today
  2. ↺ BACK TO PROPOSAL + SPECS + DESIGN

    docs(elk): record 02.26/02.27 approval + canvas instrumentation pointers

    2026-08-31 · 10 pull requests
    • #437docs(elk): record 02.26/02.27 approval + canvas instrumentation pointers
    • #438docs(elk): tee up apply artifacts for the kit-generation slice
    • #439feat(elk): wording pipeline foundations (3.5a/b/d)
    • #441feat(elk): scene ladder — ten kit images + template (3.5c)
    • #442feat(elk): fulfillment builds the kit from the listing (3.4b/3.5e)
    • #444feat(elk): /check route, CTA switch, generating state (3.4a/c, 3.9)
    • #445feat(elk): de-niche result + emails; promote founder copy rules (3.6/3.8)
    • #446feat(elk): fully-built refresh cards + resolution chips (3.2/3.3)
    • #447test(elk): evaluate-route smoke — PII, cache, throttle (4.2/4.4)
    • #448feat(elk): compose kit text with the existing OpenAI key

History read from a manifest built at 2026-09-30 (09e7bcc).

14 pull requests, found by commits touching the change folder