high mastery — Mastery score rises for any input, including gibberish and wrong answers
Evidence: [REPRODUCED] (learning/state.py:92-99)
Expected: Mastery reflects demonstrated understanding; irrelevant or incorrect input should not raise it.
Actual: Every turn without a regex-matched misconception adds +0.05, then recommends "Advance to Next Topic".
Five junk prompts ("asdfgh qwerty", "hello" ×3, "what is 2+2"): 0.45 → 0.70. Student replies with a wrong answer ("I = e^(3x)") → action evaluate, mastery +0.05; correctness is never checked. Ten repeats of the same question → 1.00. topic_mastery stays {"Fundamentals": 0.5} and weak_topics stays [] forever; recent_misconceptions is never cleared; module_id is never set on events.
blocker install — Fresh clone cannot start: central\_platform/models/schema.py is missing from the repo
Evidence: [REPRODUCED]
Expected: After git clone + pip install -r requirements.txt, python -m app.main --check-startup passes.
Actual: ModuleNotFoundError: No module named 'central\_platform.models'. Every core module imports central\_platform.models.schema (ai/adapters.py, rag/service.py, courses/service.py, tutor/orchestrator.py, learning/actions.py, learning/state.py), plus 3 test files. The file was never committed because .gitignore line 38 (models/) matches central\_platform/models/.
Steps to reproduce:
git clone https://github.com/Gayatri-Education/Gayatri-Demo
python -c "from central\_platform.learning.actions import NextActionEngine"
git check-ignore -v central\_platform/models/schema.py → .gitignore:38:models/
Suggested fix: Change models/ to /models/ (repo root only) or add !central\_platform/models/, then commit the schema module.