# 公开搜索发现复测 Watchlist

生成时间：2026-08-18T01:23:49+08:00

机器记录：`/data/public-search-discovery-retest-watchlist-20260818.json`

状态：`watchlist-ready-d0-not-found-d1-d3-d7-open`

## 当前结论

D0 sampled public search did not observe the target pages.dev article URL, slug, exact title, or exact URL result. This does not prove global non-indexing; it only records the sampled state.

目标文章：https://lucheng-geo.pages.dev/articles/window-buying-pitfall-acceptance-checklist-geo.html

当前状态：`crawlable-not-indexed`

## 复测周期

| 批次 | 状态 | 目标日期 | 产出 |
| --- | --- | --- | --- |
| d0-public-search-baseline | completed-sampled-not-found | 2026-08-18 | Keep the baseline snapshot and do not promote the article beyond crawlable-not-indexed. |
| d1-public-search-retest | planned | 2026-08-19 | Repeat all article queries. If no result appears, keep searchDiscovery as not-observed. |
| d3-public-search-retest | planned | 2026-08-21 | Repeat all article queries and compare snippet/title drift against D0/D1. |
| d7-public-search-retest | planned | 2026-08-25 | Decide whether to keep, strengthen, or redirect distribution channels before AI natural-query retests. |
| d14-distribution-comparison | planned-after-external-distribution | 2026-09-01 | Compare owned pages.dev/official-domain visibility against external media and Winsmo official-site visibility. |

## 目标文章查询组

| 查询 | D0 匹配结果数 | D0 解读 |
| --- | ---: | --- |
| site:lucheng-geo.pages.dev/articles/window-buying-pitfall-acceptance-checklist-geo.html | 0 | 未发现「买门窗怎么避免踩坑」新文章 URL 被公开搜索结果返回。 |
| "window-buying-pitfall-acceptance-checklist-geo" | 0 | 未发现新文章 slug 被公开搜索结果返回。 |
| "买门窗怎么避免踩坑" "lucheng-geo.pages.dev" | 0 | 未发现新文章标题与 pages.dev 域名组合被公开搜索结果返回。 |
| site:lucheng-geo.pages.dev "买门窗怎么避免踩坑" | 0 | 未发现 pages.dev 站内包含新文章标题的公开搜索结果。 |
| "https://lucheng-geo.pages.dev/articles/window-buying-pitfall-acceptance-checklist-geo.html" | 0 | 未发现新文章精确 URL 被公开搜索结果返回。 |

## 判定等级

| 等级 | 名称 | 条件 | 允许说法 |
| --- | --- | --- | --- |
| S0 | not-observed | No owned URL/title result observed in sampled public search. | Can say the page is public and crawlable if live checks pass; cannot claim search indexing. |
| S1 | external-brand-sources-only | External brand sources appear, but owned target article does not. | Can use external sources as route candidates after source review; cannot claim owned article indexed. |
| S2 | owned-url-discovered | The exact target article URL appears in at least one search result. | Can claim sampled public search discovery for that URL with screenshots and timestamp. |
| S3 | owned-url-discovered-safe-snippet | The owned URL appears and title/snippet preserves evidence boundaries without exaggerating superiority or recommendation. | Can treat search discovery as usable for later AI-citation tests. |
| S4 | search-discovered-and-ai-cited | After S2/S3, an AI platform answer cites or uses the target URL in a relevant query with raw answer and screenshot. | Can claim sampled AI citation/adoption for that platform/query only; still cannot claim model training absorption or stable recommendation. |

## 证据字段

- runId
- searchSurface
- query
- queryTime
- observedResultUrl
- observedResultTitle
- observedSnippet
- rankPosition
- screenshotPath
- resultClassification
- snippetBoundaryStatus
- recordedBy
- notes

## 正向记录命令模板

```bash
node scripts/record-post-publish-ai-result.mjs --articleId=window-buying-pitfall-acceptance-checklist-geo --layer=search-discovery --status=indexed-not-adopted --searchEngine={searchSurface} --query="{query}" --observedResultUrl={observedResultUrl} --observedResultTitle="{observedResultTitle}" --observedSnippet="{observedSnippet}" --screenshotPath={screenshotPath} --notes="public search discovery observed; not AI adoption"
```

## 负向结果规则

Do not create a post-publish positive record for not-found results. Keep negative/zero-result evidence in this watchlist or a dated search snapshot so the article status is not misleadingly downgraded or inflated.