Fixed issue where AI was generating inconsistent subtask IDs (101-105, 601-603)
instead of sequential numbering (1, 2, 3...) after the generateObject migration.
Changes:
- Updated all expand-task prompt variants with forceful "CRITICAL" instructions
- Made ID requirements explicit with examples: id={{nextSubtaskId}}, id={{nextSubtaskId}}+1
- Added warning not to use parent task ID in subtask numbering
- Removed parseSubtasksFromText post-processing that was overwriting AI-generated IDs
This ensures subtasks display correctly as X.1, X.2, X.3 format and the
`tm show X.Y` command works as expected.