diff --git a/CRITICAL_BUG_FIX_DELETE.md b/CRITICAL_BUG_FIX_DELETE.md
new file mode 100644
index 0000000..f0eec70
--- /dev/null
+++ b/CRITICAL_BUG_FIX_DELETE.md
@@ -0,0 +1,77 @@
+# CRITICAL BUG FIX - DeleteActionMapping Cascade Deletion
+
+## Date: 2025-12-16
+
+## Severity: CRITICAL - Data Loss
+
+## Summary
+DeleteActionMapping operation caused cascade deletion of ~54 action mappings during testing, reducing total from ~60 to only 6 mappings.
+
+## Root Cause
+When deleting multiple action mappings, IDs shift after each deletion. Deleting in ascending order causes wrong mappings to be deleted.
+
+### Example of the Bug:
+```
+Original mappings: #1, #2, #3, #4, #5
+Want to delete: #3, #4, #5
+
+Delete #3 → Mappings become: #1, #2, #3(was 4), #4(was 5)
+Delete #4 → Deletes what was originally #5! ✗
+Delete #5 → Deletes wrong mapping! ✗
+```
+
+## The Fix
+
+**Always delete in REVERSE order (highest ID first):**
+
+### WRONG (causes cascade deletion):
+```python
+for mapping in mappings_to_delete:
+    delete_action_mapping(mapping['id'])  # ✗ WRONG
+```
+
+### CORRECT:
+```python
+# Sort by ID descending
+sorted_mappings = sorted(mappings_to_delete, key=lambda x: x['id'], reverse=True)
+
+for mapping in sorted_mappings:
+    delete_action_mapping(mapping['id'])  # ✓ CORRECT
+```
+
+## Files Fixed
+- `comprehensive_crud_test.py` - Lines 436-449
+  - Added reverse sorting before deletion loop
+  - Added comment explaining why reverse order is critical
+
+## Testing Required
+Before using DeleteActionMapping in production:
+
+1. ✅ Restore configuration from backup (TestMKS_original.set)
+2. ✅ Test delete operation with fixed code
+3. ✅ Verify only intended mappings are deleted
+4. ✅ Verify count before/after matches expected delta
+
+## Impact Assessment
+- **Affected Environment**: Development/Test only
+- **Production Impact**: NONE (bug caught before production deployment)
+- **Data Loss**: ~54 test action mappings (recoverable from backup)
+
+## Prevention Measures
+1. **Code Review**: All delete-by-index operations must be reviewed
+2. **Testing**: Always verify delete operations with read-after-delete
+3. **Documentation**: Add warning comment to DeleteActionMapping implementation
+4. **Safe Delete**: Consider adding bulk delete method that handles ordering automatically
+
+## Related Code
+- SDK Bridge: `ConfigurationServiceImplementation.cs` - DeleteActionMapping method
+- Python Test: `comprehensive_crud_test.py` - Lines 436-449
+- Server Manager: `server_manager.py` - delete_action_mapping function
+
+## Status
+- [x] Bug identified
+- [x] Root cause analyzed
+- [x] Fix implemented in test code
+- [ ] SDK Bridge bulk delete helper (future enhancement)
+- [ ] Test with restored configuration
+- [ ] Verify fix works correctly
diff --git a/SERVER_CRUD_IMPLEMENTATION.md b/SERVER_CRUD_IMPLEMENTATION.md
index b270377..7fad503 100644
--- a/SERVER_CRUD_IMPLEMENTATION.md
+++ b/SERVER_CRUD_IMPLEMENTATION.md
@@ -167,6 +167,12 @@ message ServerData {
 - **Root Cause**: Using int32 instead of bool type
 - **Solution**: Use proper bool type as documented above
 
+**CRITICAL Issue**: Cascade deletion when deleting multiple action mappings
+- **Root Cause**: Deleting in ascending order causes IDs to shift, deleting wrong mappings
+- **Solution**: Always delete in REVERSE order (highest ID first)
+- **Status**: FIXED in comprehensive_crud_test.py (2025-12-16)
+- **Details**: See CRITICAL_BUG_FIX_DELETE.md
+
 ## Action Mapping CRUD
 
 Action mappings can also be managed via the same ConfigurationService.