CtrlK
BlogDocsLog inGet started
Tessl Logo
PluginPublic

adobe/stardust

Redesign an existing website to make it better. Built on top of impeccable.

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

Overview
Quality
Evals
Security
Files

implementation.mdskills/spec/reference/

Implementation layer

judgement/implementation.json says how each part must be built on EDS. Start from templates/implementation.json.

features[]

{ id, class, name, evidence, disposition, reproducibility, pattern, eds, decisions[], reach } — class, disposition, reproducibility and status use the dynamics taxonomy (../../dynamics/reference/classes-and-signals.md, triage.md); S10 stops on any other value. Run the dynamics detector on the variant representatives (≥ 80% of each template) plus one page per data-driven block, widget and form type, then curate its findings into features. reach selects the pages, in the rule format (knowledge.md § Rules), e.g. live sitemap block:news-list,event-list or live sitemap signal:iframe:*forms*. signals in the same file adds regex signals over the raw HTML ({ "hcp": "HCP_CONTENT" }). A feature on ≥ 90% of live sitemap pages is site-wide.

Martech (spec-martech output → features, vendors, launch rules, data layer)

  • Tag manager: rules matching page paths with .html, or DOM selectors, break silently when URLs and markup change — they are the rewrite sheet. Campaign rules whose path no longer resolves are retire candidates.
  • Data layer: fields read by data elements must be emitted from page metadata; data elements that read the DOM ($(".selector")) break on new markup — list them.
  • Consent: record the geo rule sets (model, default state per category) in consent_summary; the CMP loads before the tag manager; tags stay gated by consent category.
  • loading_order: head (CMP), eager (data layer), delayed (tag manager), on interaction (heavy vendors).

metadata_contract[], query_indexes[], site_config[], template_labels{}

  • metadata: { name, source: "body:data-x" | "meta:<name>", aem, used_by[] } — coverage is measured at build.
  • indexes: { name, include[], exclude[], filter, properties[], consumers[], source } — the build drafts helix-query.yaml.
  • site_config: CDN, CSP and security headers, robots/sitemaps, URL format, error pages, structured data.

Locales

Inventory every sitemap (spec-inventory counts them). For multi-language sites, write i18n_notes: tree structure, inheritance (live copies / masters), hreflang presence, what changes per tree (strings, search index, consent, gates). Fetch two to four trees fully (S2–S3 with another scopePath) when coverage of the block set across trees matters.

open_questions[]

Start from templates/open-questions.json. Each: { id, area, owner: "stakeholder" | "implementer", blocking, question, context, options[], default, default_option, impact, link }; blocking and default_option as in knowledge.md § Files. impact: a number rule (knowledge.md § Rules), e.g. urls sitemap flag=empty, or value:<n>.

skills

README.md

tile.json