| Titre | ghantakiran splunk-mcp-integration 0b86b09 Path Traversal |
|---|
| Description | Although the original generic report focused on download endpoints, the actionable issue in this repository is earlier in the export pipeline.
create_csv_export() accepts user-controlled job_name and forwards it to csv_generator.generate_csv(). The generator constructs:
file_name = f"{job_name.replace(' ', '_')}_{job_id}_{int(time.time())}.{file_extension}"
file_path = os.path.join(settings.CSV_OUTPUT_DIR, file_name)
Only spaces are replaced. Forward slashes, backslashes, and traversal tokens survive unchanged, so a job_name like ../../../../tmp/csv_poc causes the background worker to create and write a CSV outside the configured export directory. |
|---|
| La source | ⚠️ https://github.com/ghantakiran/splunk-mcp-integration/issues/49 |
|---|
| Utilisateur | LargeW (UID 97302) |
|---|
| Soumission | 14/04/2026 14:38 (il y a 6 mois) |
|---|
| Modérer | 01/05/2026 11:32 (17 days later) |
|---|
| Statut | Accepté |
|---|
| Entrée VulDB | 360542 [ghantakiran splunk-mcp-integration CSV Export csv_export.py create_csv_export job_name directory traversal] |
|---|
| Points | 20 |
|---|