Skip to content

Add /course/test governed-or-instructed test + Sovereign no-beacon guard - #83

Open
erinstanley358 wants to merge 1 commit into
mainfrom
feat/course-governance-test
Open

Add /course/test governed-or-instructed test + Sovereign no-beacon guard#83
erinstanley358 wants to merge 1 commit into
mainfrom
feat/course-governance-test

Conversation

@erinstanley358

Copy link
Copy Markdown
Collaborator

What

A new free, account-free diagnostic at /course/test - "Is your AI governed, or just instructed?" - plus a build guard that keeps the course pages' privacy promise honest. Top-of-funnel for the free AI-governance course; the free thing stays free.

Shaped by three review panels (campaign, MAC interface, fresh-eyes) and verified live in a browser at each stage.

Changes

src/pages/course/test.astro (new)

  • Honestly-scored 5-question test. Three real bands, including "you have real governance in place" - it does not assume you fail, and the low band reads "a starting point, not a failing grade."
  • Reports two dimensions separately - whether the limit holds (Q1-3) and whether the person the agent acts on has recourse (Q4-5) - instead of flattening them into one number.
  • Built on SovereignLayout: no account, no email wall, no tracking; scored entirely in the browser.
  • Save-as-PDF one-pager via a print-only rubric element (real markup, selectable and AT-readable).
  • Accessibility: radio fieldsets with legends, focus moves to the result, aria-live, prefers-reduced-motion respected, color-blind-safe verdict states (glyph + word + dashed cue, never color alone).

src/pages/course/index.astro

  • The "show the blind spot" demo button no longer overwrites the reader's typed text.
  • Voice, color-blind, and reduced-motion polish.

scripts/check-sovereign-no-beacons.sh (new) + package.json

  • Build guard: fails if any page carrying the Sovereign Article Protocol promise ("no beacons") ships an analytics or insights beacon. Wired into npm run build; also npm run test:sovereign.
  • Verified both directions: passes clean (12 Sovereign pages, 0 beacons) and fails on an injected beacon.

Verification

  • Full astro build passes; guard runs as the tail of the build.
  • Browser-driven: all three scoring bands, the two-dimension breakdown, the incomplete-answer guard, and reset all confirmed; zero console errors.
  • Sovereign no-tracking claim confirmed true (SovereignLayout loads no analytics, unlike Base/Elevation layouts).

🤖 Generated with Claude Code

@vercel

vercel Bot commented Sep 1, 2026

Copy link
Copy Markdown

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated
evoke Error Error Sep 1, 2026 2:18pm UTC

Shaped by three review panels (campaign, MAC interface, fresh-eyes).

New /course/test: an honestly-scored 5-question test that reports two
dimensions separately - whether the limit holds (Q1-3) and whether the
person the agent acts on has recourse (Q4-5) - instead of flattening them
into one number. No email wall, no tracking; scored in the browser. A
print-only rubric turns Save-as-PDF into a clean one-pager.

/course: the "show the blind spot" button no longer overwrites the
reader's typed text, plus voice, color-blind, and reduced-motion polish.

scripts/check-sovereign-no-beacons.sh: fails the build if any page
carrying the Sovereign Article Protocol promise ships an analytics or
insights beacon. Wired into `npm run build`. Verified it passes clean
(12 pages, 0 beacons) and fails on an injected beacon.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QAb6FeJoiY6JFiuCjNMfwv
@erinstanley358
erinstanley358 force-pushed the feat/course-governance-test branch from 62e1c94 to dde6e95 Compare September 1, 2026 14:18
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant