Add BreadcrumbList schema to compliance, features and guides

[vpetersson-bot]

What the report showed

In the 2026-08-04 Semrush structured-data export, Breadcrumb items = 0 for every /compliance/, /features/ and /guides/ URL, while blog posts, case studies and FAQ pages all report 1.

That matches the code: layouts/partials/head.html had dedicated BreadcrumbList blocks for blog posts, case studies, the FAQ index and individual FAQs — and nothing for the three other content trees. No page in them emitted breadcrumbs.

The crawl only sampled part of each section. In the current build it is 42 pages, not the ~20 visible in the CSV.

The change

One generic block driven by a section -> label dict, rather than three near-identical copies of the existing pattern. Section indexes get a two-level trail (Home > Guides); child pages get three (Home > Guides > Rust).

Verification

Built with hugo --minify --environment production and parsed all 479 output pages:

Section	Pages with BreadcrumbList
compliance	16/16 content pages
features	6/6 content pages
guides	20/20 content pages

Three URLs under those sections are not covered, correctly: /compliance/ntia/, /features/public-sbom-portal/ and /features/sbom-hub/ are alias stubs that emit rel=canonical + meta http-equiv=refresh. They should not carry breadcrumbs.

Also checked across the whole build:

• No invalid JSON-LD in any of the 479 pages.
• No page emits a duplicate BreadcrumbList. The existing blog / case-study / FAQ blocks are untouched — none of those sections appear in the dict, so there is no overlap.

bun run lint and the production Hugo build both pass.

Deliberately not changed

Two other columns in the report looked like findings but are not defects, so I left them alone:

• Logo items = 0 on every page. The Organization schema already includes a valid logo ImageObject with url/width/height. This is a Semrush reporting quirk.
• Product snippet items = 1–2 on most pages. Parsing the build shows exactly one Product entity site-wide, on the homepage, which is correct. Semrush appears to be counting nested entities.

The report also flags trust.sbomify.com pages as having Twitter Cards = 0 while sbomify.com reports 1. That is a real bug, but it lives in the webapp repo, not here — the app's base templates emit property="twitter:*" where the Twitter Card spec requires name="twitter:*". Out of scope for this PR; worth a separate one.

🤖 Generated with Claude Code

https://claude.ai/code/session_01FbqYj6ztzG6yautDpRWDXS