fix(scorm): mark Welcome page unmarkForCompletion so LMS completion reports

[RichardoC]

The 0.0.3 release reported 0% completion to an LMS even
though the learner finished every quiz. Root cause, traced through the
HAR of a real attempt plus Xerte's xttracking_scorm1.2.js + xenith.js:
with navigation="Menu with Page Controls", Xerte inserts a built-in
Table-of-Contents page as x_pages[0] AFTER toCompletePages is built
from currActPage over the 43 content nodes. So every content page's
page_nr (used by exitInteraction to mark completedPages[i]) is
currActPage + 1 -- a systematic off-by-one:

• toCompletePages[0] == 0 could only be matched by page_nr == 0 (the
  menu), which is never exited (menu pages are excluded from
  x_endPageTracking), so completedPages[0] (the first content page,
  "Welcome") stayed false forever.
• getSuccessStatus() returns 'incomplete' whenever any completedPages[i]
  is false, so finishTracking() set cmi.core.lesson_status='incomplete'
  (a no-op against Moodle's pre-set 'incomplete'), and the LMS never
  left the incomplete state. cmi.core.score.raw WAS being sent (75, then
  100) but never mattered to the displayed completion.

Changes to source/data.xml + source/preview.xml (identical):

• unmarkForCompletion="true" on the Welcome page
  (linkID="PG1781882348053"). This makes markedPages skip currActPage
  == 0, so toCompletePages becomes [1,2,...,42], aligning with the
  content pages' page_nr values. completedPages[0] (Welcome) is now
  marked true when the learner leaves it.

Validated end-to-end through a running XOT (docker-container branch)
instance on host port 8088:

• Pushed source/data.xml + source/preview.xml into the container; both
  carry unmarkForCompletion="true" on the Welcome page and are
  byte-identical to the repo source.
• play.php?template_id=1 returns 200; HTML5 editor opens.
• SCORM export is valid SCORM 1.2 (one SCO, 13 MB) and the exported
  template.xml carries unmarkForCompletion="true", trackingMode="full",
  trackingWeight 1x7 + 21, trackingPassed="80%", 45 questions, 8
  quizzes, 0 empty options -- all unchanged from 0.0.3 except the fix.
• Drove the exported package against a SCORM 1.2 mock API: a completed
  run sends cmi.core.lesson_status="passed" (score.raw=100), a failing
  run sends lesson_status="failed" (score.raw=62.5), and an incomplete
  run sends lesson_status="incomplete" + cmi.core.exit="suspend".
  Before the fix the completed run reported "incomplete".

Trade-off: the Final quiz (page_nr == 43) is still outside
toCompletePages (no index 43), so its completion is not tracked in
completedPages -- but its score is still tracked via its interactions
and trackingWeight="21", so the LMS grade and pass/fail are correct.
The Welcome page is no longer "required" for completion, which is
acceptable for an intro page. A proper fix would be in the Xerte engine
(xenith.js: build toCompletePages from page_nr after the menu is
inserted), but that would not survive XOT rebuilds; this content-level
workaround is used because the release workflow builds from the XOT
engine. Do not remove unmarkForCompletion="true" from the Welcome page
without a replacement fix -- re-introducing the off-by-one will silently
break LMS completion reporting again.

Guidance updates:

• docs/PROJECT_CONTEXT.md: added "SCORM completion bug fix" section
  documenting the root cause, fix, and trade-off; noted the Welcome
  page carries unmarkForCompletion="true" in the current-state list.
• docs/COURSE_VERIFICATION.md: added "LMS completion fix (post-0.0.3)"
  section with the HAR evidence, root cause, fix, and validation
  results.

Generated with pi 0.80.3 and GLM 5.2

[github-actions]

📖 Course content preview

Download the course-preview artifact from the workflow run and open index.html to read the full course (pages + quizzes, correct answers marked ✓):

• Workflow run: https://github.com/RichardoC/secure-code-training/actions/runs/28938912404
• Artifact (scroll to "Artifacts"): https://github.com/RichardoC/secure-code-training/actions/runs/28938912404#artifacts

This preview is generated from source/data.xml by tools/render_preview.py. It is for review only — the playable SCORM package is built by the release workflow.