Back to submissionsProvisional result
Lecturer review required
Jim Carrey / Agentic Student Support Workflow
9 of 16 tests passed / Evidence completeness 82%
AI-generated evaluation is advisory. The lecturer reviews the evidence and makes the final academic decision.
55%
Provisional overall mark
Pass-level evidence profile
9 tests passed 7 tests failedRubric breakdown
Functional correctness
62%35% weighting / 21.7 marks
Agentic workflow and orchestration
56%20% weighting / 11.2 marks
Safety and human oversight
50%15% weighting / 7.5 marks
Testing and reliability
42%15% weighting / 6.3 marks
Code quality
58%10% weighting / 5.8 marks
Documentation
48%5% weighting / 2.4 marks
Evidence and observations
Strengths
- The basic workflow returns the required structured result.
- Issue classification works for clear extension and technical messages.
Areas for improvement
- Empty and whitespace-only inputs are not rejected.
- High-risk escalation only runs for one exact phrase.
- The test suite covers happy paths but not safety behaviour.
Code-level observations
- src/index.ts contains the full workflow in one function.
- Risk handling relies on exact string matching.
- README.md provides install commands but no design rationale.
Personalised feedback
Editable draftRecommended learning actions
- Validate trimmed input before classification.
- Use a dedicated risk-assessment function and test several high-risk phrases.