seo: roll up sitemap lastmod from newest descendant on branch pages

[workprentice]

What

Sitemap <lastmod> for section/taxonomy/term/home pages now rolls up to the newest lastmod among the pages they render, instead of reporting only the branch page's own _index.md git-commit date.

Why

Root-caused today via a live sitemap audit: /blog/ reports lastmod 2024-05-16T10:42:54-07:00 (826 days old) while its newest post, "Compatibility Testing Pulumi HCL," published 2026-08-14. Sixteen URLs share that identical stamp (/blog/, /why-pulumi/, /automation/, /careers/, and others), all touched by the same one-time bulk commit -- enableGitInfo: true sets .Lastmod from the content file's git history, and Hugo does not roll a branch page's .Lastmod up from its descendants.

That is not merely stale, it is actively wrong: /blog/ changes constantly and the sitemap is claiming otherwise. Gary Illyes confirmed (2026-07-16) that Google evaluates lastmod trustworthiness sitewide, not per-URL, and stops trusting the field across an entire site once it catches enough incorrect dates -- so a wrong date on a high-traffic index page carries a cost well beyond that one URL.

Fix

In layouts/_default/sitemap.xml, for pages whose .Kind is home, section, taxonomy, or term, compute lastmod = max(own .Lastmod, newest descendant's .Lastmod via .RegularPagesRecursive) before emitting it. Leaf pages (.Kind == "page") are untouched, so genuinely static, old-but-honest content -- /tutorials/glossary/*, /trademark/, /acceptable-use/, /awards/ -- keeps its real date. No fabricated dates: the rollup only ever surfaces a real, already-existing .Lastmod value from a real descendant page.

Verification

Built an isolated Hugo harness (minimal site, real layouts/_default/sitemap.xml template, synthetic content tree with explicit lastmod front matter) covering:

• A section with a newer and an older child -> section reports the newer child's date, each leaf keeps its own.
• Two levels of nesting -> the outer section's date rolls up transitively through the inner subsection.
• A taxonomy list page and one of its terms -> both roll up across tagged content.
• A section with zero children -> keeps its own date (no crash, no false rollup).
• Standalone leaf pages -> completely unaffected.
• Home page -> rolls up across the whole site, as expected.

All eleven pages in the harness sitemap emitted exactly the expected <lastmod> value; no XML malformed, no unexpected date on any leaf.

Scope

layouts is fully excluded from prettier (.prettierignore), so no formatting step applies. This is a template-only change with no content or front-matter edits.

---

🧠 This PR was created by workprentice on behalf of the Pulumi SEO/AEO daily optimization assignment.

[github-actions]

Pre-merge Review — Last updated 2026-08-21T12:21:45Z

Tip

Summary: Infrastructure-only change to layouts/_default/sitemap.xml: for branch pages (home, section, taxonomy, term) the emitted <lastmod> is now the newest of the page's own .Lastmod and every descendant regular page's .Lastmod, instead of the branch page's own commit date. That's the right SEO behavior — a section index whose children changed yesterday should no longer advertise a two-year-old lastmod. The template logic reads correctly: $lastmod is seeded from .Lastmod so leaf pages are unaffected, outer-scope assignment inside range is valid in modern Hugo, and the zero-time guard now tests the rolled-up value. The kind of wrongness that would matter here is silent output corruption (a malformed or missing lastmod for every URL) or a build-time regression on a site this large; I reviewed the template semantics for the former and flagged the latter as a thing to watch. No content files changed, so no prose, link, frontmatter, or code-example passes applied.

Review confidence:

Dimension	Level	Notes
mechanics	MEDIUM	Template logic reviewed by reading; CI doesn't run hugo here, so the rendered sitemap and full-build timing weren't observed.
facts	HIGH	No factual claims — the diff is template code.

Investigation log

• Cross-sibling reads: not run (not in a templated section)
• External claim verification: 0 of 1 claims verified (0 unverifiable, 0 contradicted) · 4 specialists (numerical, cross-reference, capability, framing); 0 cross-specialist corroborations · routed: 0 inline, 0 Pass 1, 0 Pass 2, 1 Pass 3 (verified 0, contradicted 0, unverifiable 1).
• Cited-claim spot-checks: not run (no cited claims)
• Frontmatter sweep: not run (no frontmatter in diff)
• Temporal-trigger sweep: not run (no trigger words)
• Code execution: not run (no static/programs/ change)
• Code-examples checks: not run (no fenced code blocks in content files)
• Editorial-balance pass: not run (not under content/blog/)

🚨 Outstanding	⚠️ Low-confidence	💡 Pre-existing	✅ Resolved
0	1	0	0

🔍 Verification trail

2 claims extracted · 0 verified · 1 unverifiable · 0 contradicted

• L13 in layouts/_default/sitemap.xml "{{ safeHTML ( $lastmod.Format '2006-01-02T15:04:05-07:00' ) }}{{ end }}{{ with .Sitemap.ChangeFreq }}" → ➖ not-a-claim (evidence: This is a Hugo template snippet formatting a lastmod date field using Go time layout syntax; it is code/template logic, not a factual or falsifiable assertion.; source: layouts/_default/sitemap.xml)
• L7 in layouts/_default/sitemap.xml "nested RegularPagesRecursive walk runs once per branch page, including the home page" → 🤷 unverifiable (evidence: build-time cost depends on total page count on a full registry build, which this review environment does not run; source: reading of the template's loop nesting)

🚨 Outstanding in this PR

No outstanding findings in this PR.

⚠️ Low-confidence

Review each and resolve as appropriate — these don't block the PR.

• [L7] layouts/_default/sitemap.xml — the roll-up nests a RegularPagesRecursive walk inside the .Data.Pages loop, so on a full build every branch page re-walks its entire subtree, and the home page walks the whole site. On a small site that's free; on a full registry build it's roughly the sum of every branch page's descendant count. Flagged 🤷 unverifiable because CI doesn't run hugo for this job, so I couldn't measure it. Worth a glance at the sitemap-generation time in the next full build before merge; if it moves, the usual fix is to compute the roll-up once per section rather than per page in the output loop.
  • Two smaller things to consider while you're in here, neither blocking: the walk includes descendants that the if not (or .Params.private ...) guard excludes from the sitemap, so a private or block_external_search_index page can still bump its parent's lastmod; and because only regular pages are walked, a nested _index.md edited on its own won't propagate up to its ancestors. Both are defensible as-is — the first arguably reflects real content churn — but they're the two places the emitted date won't match what a reader would guess.

💡 Pre-existing issues in touched files (optional)

No pre-existing issues in touched files.

✅ Resolved since last review

No items resolved since the last review.

📜 Review history

• 2026-08-21T12:21:45Z — Sitemap lastmod roll-up reads correct; no blockers, one note on the nested per-branch-page walk's build cost (9b4253d)

---

• Refresh this review — comment @claude #update-review. Say what you fixed, or which finding you dispute and why; both work in the same mention.
• Ask for anything else — comment @claude with no hashtag (questions, one-off fixes). Leaves this review untouched.

Important

Please don't hide, resolve, or delete this comment! It breaks things!

📖 How pre-merge review works — the full lifecycle, short-circuits, and escape hatches.

[pulumi-bot]

Your site preview for commit 9b4253d is ready! 🎉

http://www-testing-pulumi-docs-origin-pr-21036-9b4253d4.s3-website.us-west-2.amazonaws.com

[pulumi-bot]

Lighthouse Performance Report

Commit: 9b4253d | Metric definitions

Page	Device	Score	FCP	LCP	TBT	CLS	SI
Homepage	Mobile	🔴 32	3.0s	8.2s	3537ms	0.072	8.4s
Homepage	Desktop	🟢 92	1.2s	1.5s	0ms	0.012	1.2s
Install Pulumi	Mobile	🟡 57	7.6s	16.1s	25ms	0.000	7.6s
Install Pulumi	Desktop	🟢 92	1.2s	1.5s	0ms	0.022	1.2s
AWS Get Started	Mobile	🟡 60	5.4s	7.2s	2ms	0.102	5.4s
AWS Get Started	Desktop	🟢 92	1.1s	1.5s	0ms	0.071	1.1s