1. Report prompt now extracts user's core questions from simulation_requirement and structures the report to directly answer them (e.g. "will it go well?", "where are the problems?") instead of generic future-prediction questions. 2. Step3Simulation auto-triggers report generation 1.5s after simulation completes. 3. ReportView auto-redirects to latest successful report when current report has failed, via /api/report/check/<simulation_id> endpoint. |
||
|---|---|---|
| .. | ||
| app | ||
| scripts | ||
| pyproject.toml | ||
| requirements.txt | ||
| run.py | ||
| uv.lock | ||