POST /api/assessments/:definitionId/start only ever checked that the
definition existed and was active — it never checked unlock_rule. GET
/api/capstone correctly computed and displayed lock state, and the
student UI never exposed a start action while locked, but a direct POST
to this route bypassed the check entirely (documented as a known gap in
docs/PHASE6A_ROOT_COURSE_FOUNDATION.md §10). Student UI state is not the
enforcement boundary.
Fixed by binding to the exact same computeCurriculumState() gate
/api/capstone already relies on, rather than re-parsing
definition.unlock_rule independently — one computation of 'is this
unlocked,' not two that could drift. Only a definition the curriculum
path actually names as ITS capstone is gated this way; an ordinary exam
is completely unaffected. Returns a clean 403 { error: 'locked',
blocking_stage_ids } — the same shape /api/capstone already exposes.
Found while fixing this: e2e-phase5b.js's own capstone-completion test
had been passing only because of this exact gap — it never included the
real-ops stage in its completion loop before attempting to start the
capstone, and the missing unlock check let it succeed anyway. Fixed
alongside the platform code (REAL_OPS added to ALL_NON_CAPSTONE), not
routed around.
Co-Authored-By: Claude Sonnet 5 <[email protected]>
New learning-routes.js: scenario start/attempt-state/hint (sequential reveal, 409 once
exhausted, policy-aware under an active assessment), GET /api/curriculum/state (rule-driven,
alongside the unchanged /api/curriculum/path), assessments list/start/active/detail/submit,
GET /api/capstone, GET /api/achievements, GET /api/attempts/recent.
/api/progress/record now routes through attempts.recordValidation instead of a flat insert,
and, on a pass, checks curriculum/capstone achievement consequences and, if the scenario is
part of an active assessment, updates that session (assessment context is derived entirely
server-side from assessment_session_scenarios, never from anything the client asserts).
Gateway changes: two new relay routes (/lab/:token/api/scenarios/:id/start and .../hint,
mirroring the existing validate/answer pattern - pure Platform API operations, the Lab has no
attempt/hint concept) and a response rewrite on GET /lab/:token/api/scenarios/:id that strips
the Lab's own hints array (replacing it with a bare hint_count) so progressive reveal is
actually server-gated instead of just UI-collapsed - the Lab image's bytes stay untouched;
standalone/single-container dev mode never goes through the Gateway and keeps the full array
unaffected.
ced4568andClaude Sonnet 5