Executable quality bar

QA contract agents must satisfy

Run npm run qa:all before presenting the site as ready. Use /qa-contract.json when a machine-readable contract is preferred.

npm run qa:routes

  • Core HTML routes return 200
  • Machine endpoints return 200
  • Internal links return 200
  • Sitemap URLs resolve

npm run qa:seo

  • Unique titles
  • Valid descriptions
  • Canonical URLs
  • One H1
  • Open Graph parity
  • Favicon
  • JSON-LD

npm run qa:aeo

  • llms.txt
  • llms-full.txt
  • schemamap.xml
  • Schema endpoint JSON
  • Contract registry
  • Site studies
  • SEO/AEO evaluation
  • Source traceability
  • Organization/WebSite facts
  • FAQ/LocalBusiness where expected

npm run qa:browser

  • Desktop screenshots
  • Mobile screenshots
  • Visible H1
  • Horizontal overflow
  • Viewport overflow
  • Potential text clipping report
  • Landmark/name/alt/duplicate-id/focus heuristics
  • Screenshot baseline visual diff with mutation control
  • Keyboard focus navigation per desktop route
  • axe-core WCAG 2.2 AA accessibility scan (2.0/2.1/2.2 tag set)

npm run qa:crawl

  • Required crawl inventory exports
  • Screaming Frog issue export summary
  • Critical response-code findings
  • Crawl evidence report under reports/crawl/

npm run qa:agent

  • Main URL orientation simulation
  • SEO/AEO walkthrough selection
  • Crawl/browser readiness simulation
  • Handoff continuation simulation

npm run qa:contracts

  • Live contract registry endpoints
  • Schema/source existence
  • Missing endpoint fixture
  • Missing schema fixture
  • Missing source fixture
  • Missing evidence fixture
  • Execution-track document statuses match /agent-tasks.json
  • Execution-track drift fixtures fail as expected

npm run qa:copy

  • Weak copy fixtures fail control checks
  • Improved copy includes required conversion signals
  • Required principles map to the copy contract
  • Forbidden phrases do not survive transformation
  • Report saved under reports/copy-stress/

npm run qa:crawl-stress

  • Static export re-crawled from homepage, sitemap, and manifest seeds
  • Broken internal links fail with referrers
  • Redirect chains beyond budget fail
  • Documented response-time and response-size budgets enforced
  • Sitemap URLs unreachable from the homepage fail
  • Synthetic negative controls fail as expected

npm run qa:playbooks

  • Every site-type playbook has a simulated build fixture
  • Required inputs are provided or explicitly waived
  • Planned pages map to known page patterns and required schema
  • Published claims carry simulated source IDs and blocked claims carry reasons
  • Playbook QA gates and stress tests are covered
  • Unresolved risks are declared
  • Negative controls fail in expected ways

npm run qa:perf

  • Local LCP/CLS measured per route against budgets
  • HTML/CSS/JS/image/font asset budgets enforced
  • Images require dimensions; head scripts may not block; stylesheet count capped
  • Zero-JS default with explicit exception list
  • Synthetic budget-violation controls fail as expected

npm run qa:demo

  • Demo source chain files exist
  • Demo pages implement the full playbook page model
  • Required schema types render in JSON-LD
  • Visible claims carry allowed ledger claim IDs
  • Blocked phrases stay off pages, schema, and LLM surfaces
  • Fictional boundary stays visible
  • Negative controls fail as expected

npm run qa:domain

  • Local static route readiness
  • Machine route readiness
  • Required security headers configured in public/_headers
  • Optional QA_PRODUCTION_URL apex/www/http/https/TLS variant evidence
  • Optional QA_PRODUCTION_URL live security-header verification

npm run qa:nectiv

  • LLM surfaces for AI/AEO claims
  • Indexable H1/description/canonical guardrails
  • Sitemap/core-route parity
  • Crawler-visible critical URL findings
  • Production-domain evidence warning

npm run qa:handoff

  • Generated handoff report
  • QA evidence summaries
  • Residual risks
  • Next actions from /agent-tasks.json

npm run qa:all

  • Spec audit
  • Astro check
  • Build
  • Route QA
  • SEO QA
  • Browser visual QA
  • Performance budget QA
  • Crawl evidence QA
  • Crawl stress QA
  • AEO QA
  • Agent simulation QA
  • Contract drift QA
  • Copy stress QA
  • Page-pattern fixture QA
  • Playbook simulation QA
  • Demo build QA
  • Domain/static QA
  • Generated handoff QA