tools/llm-oneshot/apps/paint-app/prompts/grading_rubric.md
Each feature: 0–3 points
| Score | Meaning |
|---|---|
| 0 | Not implemented or broken |
| 1 | Partial - major issues |
| 2 | Mostly working - minor bugs |
| 3 | Fully working |
| Feature | Max | Score |
|---|---|---|
| 1. Basic Drawing | 3 | |
| 2. Live Cursors | 3 | |
| 3. Shapes | 3 | |
| 4. Selection | 3 | |
| 5. Layers + Locking | 3 | |
| 6. Presence | 3 | |
| 7. Comments | 3 | |
| 8. Version History | 3 | |
| 9. Permissions | 3 | |
| 10. Follow Mode | 3 | |
| 11. Activity Feed | 3 | |
| 12. Sharing | 3 | |
| 13. Canvas Chat | 3 | |
| 14. Auto-Cleanup | 3 | |
| 15. Text & Stickies | 3 | |
| TOTAL | 45 |
These criteria specifically test SpacetimeDB strengths (bold items above):
Score these separately to measure STDB advantage.