docs(blog): publish paired English editions and complete indexes

[huangruiteng]

Goal And Delivered Outcome

• Goal/source and gap: The public English Blog index exposed only the original architecture article even though two newer Chinese articles were already published. The owner requested complete English editions and a durable, fast test that prevents later translation or index drift.
• Observable before → after, with the validation row that proves it: /blog/ listed one post and the two corresponding English routes did not exist; it now lists all three posts, both newer articles have complete English pages, English diagrams or interaction copy, reciprocal language navigation, and canonical alternates. The focused source test and share-bundle integration row below dynamically discover and verify the bilingual catalog.
• Issue/task and intended base: Direct owner request; base main at 092a28c4e7704828b134e21757edb29924ff9b03.

Scope And Continuation

• Completed scope and remaining work: Complete within this scope. Published the English application-scenarios and agent-native Kanban editions, added four English Kanban diagrams, completed both Blog indexes and language switches, and replaced the hard-coded Blog inventory with a shared filesystem-discovered validator used by a fast source test and the packaged-site smoke.
• Slice boundary / successor: N/A. The existing static-site owner, Vite export path, shared editorial styles, and Frontstage Pages workflow remain the owning boundary. The future-facing pass was applied by centralizing bilingual catalog validation instead of adding another manual allowlist.

Validation

• Tested revision: 163cc3cc7d9e06ad2cde7da2a1ee09944a7ebcae
• Run state: finished
• Input classes: synthetic, public_fixture

Check kind	Result	Public-safe evidence / limitation
unit	passed	node examples/blog-bilingual-index-smoke.mjs; fast source-level discovery verifies three paired articles, both locale indexes, lang, canonical URLs, hreflang, counterpart links, visible titles, and paired section identities. Wired into Frontstage Pages CI.
integration	passed	uv run --extra test node examples/frontstage-share-bundle-smoke.mjs; built the public site, reused the bilingual validator against packaged output, and verified static fallback, scripts, diagrams, and repository-base links.
static	passed	node --check for the English interaction and bilingual-check scripts, xmllint --noout for all four English SVGs, examples/frontstage-pages-workflow-smoke.py, and git diff --check origin/main...HEAD.
manual	passed	Real-browser desktop and 390 px mobile inspection covered the English index and both new article first screens; all diagram requests returned SVG 200 responses; presentation mode and all synthetic pull-request states changed visible content correctly. The owner approved the first-screen preview before commit.
regression_parity	passed	English and Chinese editions have matching section identities and matching structural counts for sections, paragraphs, list items, and tables; the focused test rejects a missing locale directory, index link, alternate, counterpart link, or section.
static	passed	Exact-head loopx canary premerge --from-git-diff --git-diff-base origin/main --tier standard: 10 catalog canaries, 8 content-ops risk checks, diff hygiene, and the public-boundary scan passed with no manual holds. The initial earlier run exposed only a missing declared TypeScript parser dependency in the clean worktree; after npm ci --ignore-scripts, the failed semantic-vocabulary check and both subsequent full gates passed without lowering coverage.

• Coverage and gaps: The changed surface is static editorial HTML, CSS-compatible SVG, one page-local progressive-enhancement script, the focused bilingual test, and packaged export validation. Desktop/mobile rendering, JavaScript and no-SPA export behavior, locale structure, assets, links, workflow wiring, and the public/private boundary are covered. Runtime and real-backend validation are not applicable.

Frontend / Visual Evidence

• UI impact: changed
• Before: The production English Blog index lists only the original architecture post; the two newer English article routes are absent.
• After: The English index lists all three posts newest-first. Both new English articles use the existing LoopX editorial system, reciprocal language switches, and English diagrams or interaction content. Public-safe desktop and mobile previews were reviewed and approved by the owner before commit; generated review screenshots were not added to the repository.
• States and viewports shown: Blog index and both article first screens at 1440×900 and 390×844; application-scenarios reading/presentation modes and four synthetic PR states; Kanban article diagrams.
• Source data: synthetic, public_fixture

Type of Change

• Bug fix
• New feature
• Breaking change
• Refactoring (no functional changes)
• Documentation update
• Test update

LoopX Area

• Control plane (goals, todos, quota, scheduler, registry, runtime)
• Benchmark boundary (adapters, runners, verifiers, scoring, evidence)
• Capability or extension (providers, adapters, skills)
• Public docs or presentation surface (README, protocols, dashboard)
• Build, packaging, installer, or CI
• Host or runtime integration

Technical Direction

• Direction / acceptance reference, when applicable: N/A; this is a self-contained public Blog completeness repair.

Shared-authority RFC fixture impact

N/A. This PR does not claim shared Goal Authority or control-plane migration progress.

Boundary Checklist

• Neither the diff nor this PR body/comments/attachments disclose private state, credentials, raw traces or verifier output, internal links, or local machine paths (including .loopx/, .codex/goals/, and live ACTIVE_GOAL_STATE.md).
• I did not duplicate maintainer-owned benchmark work unless a maintainer split out a public issue for it.
• I kept the change scoped to the linked issue/task.
• I completed the visual evidence section for UI changes, or marked UI impact none.
• Every commit includes a DCO Signed-off-by trailer (git commit -s).

[huangruiteng]

Approval conclusion (author-owned PR; GitHub blocks formal self-approval)

审查对象：4701@163cc3cc7d9e06ad2cde7da2a1ee09944a7ebcae（已合并，合并后审计）。merge base 092a28c4e7704828b134e21757edb29924ff9b03，14 个文件 +771/-9。

动机

站点上的 Blog 是双语的：apps/presentation/site/public/blog/ 对应英文，blog/zh/ 对应中文。但在 base 上，英文索引只列了 from-one-shot-agents-to-long-horizon-control 一篇文章，agent-facing-kanban 和 application-scenarios 只有中文版；同时唯一覆盖 Blog 的 examples/frontstage-share-bundle-smoke.mjs 用的是硬编码文章清单，并且对这两篇没有译文的文章直接跳过 hreflang 断言。结果是：缺译文、缺索引项、缺语言互链这三类真实缺陷都不会让任何检查变红——手维护清单本身就是问题来源，每加一篇文章都要重新冒一次风险。这个 PR 把“目录即事实”的规则固化下来，并补上缺失的英文版本与索引项。

改动思路

一条规则、一个所有者：把“双语目录必须成对、每个 locale 索引必须列出全部文章、每篇文章必须带 canonical / 对方语言链接 / en+zh-CN+x-default 互链、成对文章的 section id 必须一致”抽成 validateBilingualBlog()，由新脚本 examples/blog-bilingual-index-smoke.mjs 导出（含 collectArticleSlugs() 从磁盘发现文章集合）。旧的 frontstage-share-bundle-smoke.mjs 不再自带清单，而是直接 import 这个函数拿文章集合，并删掉那处按文章名判断 hreflang 的特例。检查入口接进 .github/workflows/frontstage-pages.yml，同时把新脚本加进 push/PR 两条 path filter，改动内容或改动检查都会触发。

具体改动

• 新增 examples/blog-bilingual-index-smoke.mjs（87 行，无依赖）：collectArticleSlugs() 用 readdir + existsSync(<dir>/index.html) 发现文章、排序；validateBilingualBlog() 断言两个 locale 的 slug 数组完全相等、两个索引都出现 href="<slug>/"、<html lang>、rel="canonical"、对方语言链接、三种 hreflang，并逐篇比较成对文章的 <section id> 序列。
• 发布内容：新增英文版 blog/agent-facing-kanban/index.html（247 行）、blog/application-scenarios/index.html（201 行）及英文 presentation.js、四张英文 SVG 图；英文索引补上这两篇（1→3 条），两篇中文文章补上三种 hreflang 与 EN 语言切换链接。
• examples/frontstage-share-bundle-smoke.mjs：改为 import { validateBilingualBlog }，文章集合由发现结果驱动，删除原有 ["agent-facing-kanban/","application-scenarios/"] 的 hreflang 例外分支。
• .github/workflows/frontstage-pages.yml：新增 path filter 两项与 “Validate bilingual Blog catalog” 步骤（node examples/blog-bilingual-index-smoke.mjs）；apps/presentation/site/README.md 记录该不变量与两个 smoke 的职责。

对主干的风险

风险面是公开静态站点与 CI，不涉及运行时代码、LoopX 控制面、CLI/API 或持久化状态，回滚只需还原内容提交或摘掉该步骤。实测：node examples/blog-bilingual-index-smoke.mjs → ok (3 paired articles)；两处变异都按预期失败——(1) 从英文索引删掉 href="application-scenarios/" 报 Blog index must link every en article: application-scenarios；(2) 只给英文 agent-facing-kanban 插入 <section id="mutation-probe"> 报 Paired Blog article sections must match: agent-facing-kanban。两点保留意见（均非阻塞，P3）：断言基于渲染后 HTML 的子串匹配，理论上可被无关出现满足，结构性更强的部分是与成对 section id 的比较；本地跑完整 frontstage 链时卡在 examples/goal-channel-frontstage-fixture.py（PATH 上 python3 为 3.9.6，不支持 dataclass(slots=True)），该失败在 merge base 上同样复现，属既有机器环境限制，非本 PR 引入，因此这一步以合并后的 CI 结论为准。新英文页面的浏览器渲染未在本地验证，仅由静态 HTML 断言与 CI 覆盖。

我的整体评价

这是一个小而完整的收口：既补上了用户可见的缺失内容（两篇英文文章 + 两个索引项），又把导致缺陷长期潜伏的手维护清单换成了目录驱动的机器校验，并且让旧 smoke 复用同一规则而不是再存一份。迁移成本为负——删掉的清单比新增的检查更多；两处变异证明该检查确实会失败而不是空转。作为已合并精确 head 的合并后审计，证据支持通过。

English verdict: APPROVE (exact head 163cc3c)